Pwyllgor yr Economi, Seilwaith a Sgiliau - Y Bumed Senedd

Economy, Infrastructure and Skills Committee - Fifth Senedd

09/05/2018

Aelodau'r Pwyllgor a oedd yn bresennol

Committee Members in Attendance

Adam Price
David J. Rowlands
Joyce Watson
Lee Waters
Mark Isherwood
Russell George Cadeirydd y Pwyllgor
Committee Chair
Vikki Howells

Y rhai eraill a oedd yn bresennol

Others in Attendance

Catherine Phillips Uwch-gynghorwr corfforaethol, Busnes yn y Gymuned Cymru
Senior Corporate Adviser, Business in the Community Wales
Christopher Hoskins Rheolwr Masnachol SoilQuest, Agrii
Chris Hoskins, SoilQuest Commercial Manager, Agrii
Jason Llewellin J Llewellin & Co
J Llewellin & Co
Leighton Jenkins Cyfarwyddwr Cynorthwyol, CBI Cymru
Assistant Director, CBI Wales
Matt Fenech Ymgynghorydd Ymchwil ac Eiriolaeth Deallusrwydd Artiffisial, Future Advocacy
AI Research and Advocacy Consultant, Future Advocacy
Professor Calvin Jones Ysgol Fusnes Caerdydd
Cardiff Business School
Professor Simon Blackmore Pennaeth Roboteg Amaethyddol, Prifysgol Harper Adams
Head of Agricultural Robotics, Harper Adams University

Swyddogion y Senedd a oedd yn bresennol

Senedd Officials in Attendance

Abigail Phillips Ail Glerc
Second Clerk
Gareth Price Clerc
Clerk
Katy Orford Ymchwilydd
Researcher
Robert Lloyd-Williams Dirprwy Glerc
Deputy Clerk

Cofnodir y trafodion yn yr iaith y llefarwyd hwy ynddi yn y pwyllgor. Yn ogystal, cynhwysir trawsgrifiad o’r cyfieithu ar y pryd. Lle mae cyfranwyr wedi darparu cywiriadau i’w tystiolaeth, nodir y rheini yn y trawsgrifiad.

The proceedings are reported in the language in which they were spoken in the committee. In addition, a transcription of the simultaneous interpretation is included. Where contributors have supplied corrections to their evidence, these are noted in the transcript.

Dechreuodd y cyfarfod am 09:21.

The meeting began at 09:21.

1. Cyflwyniad, ymddiheuriadau, dirprwyon a datgan buddiannau
1. Introductions, apologies, substitutions and declarations of interest

Bore da. Good morning. Welcome to the Economy, Infrastructure and Skills Committee. Moving to item 1, we have apologies from Hefin David this morning, and Vikki Howells will be joining us a little bit later. Are there any declarations of interest this morning? No, there are none.

2. Egluro’r cefndir—Awtomeiddio ac Economi Cymru
2. Scene setting—Automation and the Welsh Economy

In that case, I move to item 2 with regard to our inquiry on automation and the Welsh economy, and this is the first session of three. We did have a scene-setting visit last week to Swansea University and Amazon, which I think helped to get our thoughts straight. I think those were two positive meetings last week that we had.

We've got a panel of four this morning—I'll call them experts; four experts. Perhaps if you could just introduce yourselves for the record, if I go to this room first and start on my left.

I'm Calvin Jones. I am professor of economics at the Welsh Economy Research Unit in Cardiff Business School. As the title suggests, my primary interest is in the Welsh economy and economic drivers thereof. Recently, that's included lots of technology-related subjects, and we are currently delivering a project looking at how superfast broadband can enhance, should enhance, the productivity of small and medium-sized enterprises in Wales, the European regional development fund project. So, that has lots of spin-offs into automation, air and other relevant areas.

My name's Leighton Jenkins. I'm assistant director and head of policy at the Confederation of British Industry in Wales. The main reason for our interest is the growing awareness amongst the members of the CBI around the opportunities and the risks of automation and the lack of a Government strategy in Wales to look at those strengths and look at the weaknesses. So, we decided to launch a little campaign to kind of push the Welsh Government to look at this. So, that's kind of where we are at the moment.

Thank you, Leighton. Catherine, do you want to introduce yourself?

Yes. Hello, I'm Catherine Phillips. I have two jobs. I'm a corporate adviser at Business in the Community. We are a not-for-profit that works with businesses to understand their social, environmental and economic impact and how we can work with them to transform their core products and services. As a result, I lead our work on not only how the fourth industrial revolution can be a force for good but how we anticipate the unintended consequences. We've created a report called 'A Brave New World?', which came out in late 2016, and we're working on how we integrate that now into the businesses we work with and their operations.

Good morning, and thanks for the opportunity to speak today. I am Matthew Fenech. I'm an artificial intelligence research and advocacy consultant at Future Advocacy. We are a think tank and consultancy based in London. We do work in a number of areas around global Government relations to tackle big twenty-first century challenges, and under that remit, in fact, I lead the project on artificial intelligence, which is about policy development and advocacy to maximise the opportunities and minimise the risks of artificial intelligence and other automating technologies. We have three main focus areas: AI in low and middle-income countries; AI in healthcare, on which we released a report with the Wellcome Trust just two weeks ago; and the other main area, which I'm sure we'll be talking about most today, is AI and the future of work.

09:25

Thank you, Matt. This morning's session—we've got an hour and a half. It's fairly free-flowing, but I can't quite see the detail on your face. If you want to speak, perhaps just wave or something if you want to come in on a point. But feel free just to come in as well. I think we'll treat this as informal, if you like—

—although it's all, of course, on the record this morning.

If I can ask the first question: should automation be something that we should all fear? I'm looking at Calvin.

It depends what you do. I should fear it; you perhaps less so. It's easy to get caught up in the technicalities of which tasks, occupation industries, are up for replacement, if you're thinking from the kind of threat to be expected, but you have to remember, of course, that we arguably should have automated a lot of legal 10 years ago, accountancy 15 years ago, but didn't. So, it's the social context of that technology that will determine where the threats and the resistance or resilience arises. So, we should be as concerned with that as with the impact in the kind of purely economic productivity terms.

Yes, I just wanted your reflection on whether or not Wales is particularly vulnerable. The analysis suggests that the sectors that are going to be hit first are things like manufacturing, retail, low skills, back-office functions, with repetitive tasks, and we have quite a lot of those.

How long do you want on this? And that's a serious question. Do you want two minutes or 20?

Right; two minutes. There are a number of areas that Wales concentrates on that are seen as particularly vulnerable to automation. I'm actually less concerned about manufacturing because a lot of that happened in the 1980s anyway. The stuff that's left in Wales in manufacturing is the lean stuff already. In manufacturing the issue is much more about the age of capital. So, if you talk to Tata, they won't be spending £500 million on a kind of fully automated low-carbon steelworks in Port Talbot because why would you? Because Port Talbot's doing—. So, it will vary on individual plants within manufacturing. I certainly think that middle level white-collar jobs where we do not have headquarters are particularly vulnerable. Obviously, call centre type of occupations is one of those. Even where we do have headquarters—the prime case would be Admiral. At some point, Admiral will have to go with the flow and will start to shed jobs significantly. So, yes, those issues are very pertinent to Wales in the occupational sense and the industry sector sense, but as I kind of hinted earlier, as importantly, those decisions won't be made in Wales, where the company automates, and how—they won't be made here.

Well, that was less than two minutes. [Laughter.] So, what are the headlines that you're missing out there?

Skills—if I can take it to two minutes, all the evidence that I've seen in the last two or three years, looking at this, is that there are complex impacts on skill levels. There's this argument that you hollow at the middle so that the people with reasonable education who are doing white-collar repetitive tasks are the ones who will be next hit, but generally being more skilled/educated/qualified, and particularly being adaptable, maybe partly as a result of that education—maybe not—would seem to be something that is worth while in terms of being resilient to automation and other relevant technological—.

Can I just—?

I'll come to Leighton, then I'll come to you, Matt. Leighton.

I just wanted to add to Calvin's point around automation. There is a real risk here that we don't consider the quite quick changes that can happen in certain industries, and also that these people—. So, for example, the largest three private sector employers in Wales are all retailers—Asda, Sainsbury's, Tesco. They are predominantly low-skill women who work part-time in their communities to support their families. The World Economic Forum has done a study that looked at these kinds of high-risk jobs that may face the first wave of automation, and set out a realistic skills pathway for these people to retrain, and I think it's a model worth looking at. So, they have some basic principles that say, 'How do we take this woman in Tesco who was a cashier—now that there's increasingly less call for cashiers, how do we re-skill her into a job that's easy to be re-skilled in, has comparable wages and has job security, and that isn't second on the rung for automation so that she ends up retraining again?' We are nowhere near prepared for that.  

09:30

Can I just come to Matt, then I'll come to Adam? Matt. 

So, on both the initial questions, really, on whether we should fear automation, I think the answer should be 'no', because automation is a technology like any other, and it's really value free; it depends on the decisions we take around how it's implemented. There's a real opportunity here—wouldn't it be great if more dangerous or more routine and repetitive aspects of our work could be automated, freeing us up to do potentially more rewarding, fulfilling work that involves interacting with other people? So, that's one aspect on the fear aspect. 

On the impact in terms of different sectors, certainly, what we found in our research is that the impact of AI and automation is going to vary geographically. So, our study last October looked at the impact of automation across 632 parliamentary constituencies in Great Britain. We found that although the headline figure for the risk of automation of jobs in the UK is about 30 per cent, the risk actually varies in different parliamentary constituencies from 22 to about 40 per cent. Looking at Wales, that range is about 26 to 36 per cent, and what we found is that large employers in particular areas tend to skew the risk of automation in those particular areas. So, when you look at the map of Wales and look at the constituencies that are most likely to be impacted, they are concentrated around the economic heartlands in the north and in the south, where there are these large manufacturers. 

I take the previous point that maybe a lot of automation in manufacturing has happened already, but Leighton also made a good point that the top three employers in Wales are in wholesale and retail, and quite apart from the impact in terms of cashiers of automation, there is a lot of warehousing and back-end processes in wholesale and retail, and this is something that's going on at the moment. For example, there was a recent announcement about Shop Direct, who are closing three warehouses in the north-west of England to open a smaller automated warehouse in the east midlands, with fewer jobs being taken up there. So, these are processes that are ongoing at the moment. We really are, I think, at the tipping point.  

I'll bring Catherine in in a moment; I'll just come to Adam first. Adam Price. 

I was just wondering, before we delve into the detail, whether you have, particularly Calvin—. I mean, what do you think of the arguments that Mark Carney has made that we could be looking at a post-human economy, and this could lead to the resurgence of Marxism because, effectively, if you've got a form of hypercapitalism where there is only capital because there are fewer and fewer workers needed, what does that mean in terms of our entire social and economic system?  

Under current policy paradigms, he's completely right. 

So, it depends whether you think the resurgence of Marxism is a good thing. 

Yes. We might come back to this, but the critical thing in this new potentially post-labour economy is: who owns capital? That's the only thing that matters. 

Yes. So, essentially, his reading, which if you fear the revival of communism, I suppose, would be described as apocalyptic, but, actually, there's a certain logic to what he says. 

Marx wasn't wrong in any of his predictions, apart from the bit about revolution. He predicted conglomerations emerging, the concentration of industry, massive technological strides and back to capitalism, and levels of material wealth are very much higher than he would have seen. He also predicted continuous unemployment into this future—underemployment he was certainly right about. 

Now, the problem is that the thinkers who came after Marx who were in his tradition—people like, say, Schumpeter, who we forget was a bit of a post-Marxist in his quiet times—Schumpeter talked about the role of Government as a moderating force on that, and as a redistributing force, and as a force to reign in, I won't say the excess of capitalism, but the concentration of power in particularly large private sector conglomerates. And, arguably, that's sort of what has happened in Scandinavian countries, where you have a more involved state that takes more care about how much is redistributed and to whom. That vision of social democracy, arguably, you might say is in a better position to start dealing with some of these problems, as opposed to say the US, where it's the devil take the hindmost, and the hindmost being an increasingly large rump.

09:35

So, there has to be a social and political response to these new conditions. Just on the economy side, could you just say: we've had other Schumpeterian gales of creative destruction, yet they have created jobs—there's been pain in the transition; what is different, essentially, about this phase, or this type of automation that means that it will lead to a net job-loss outcome in your view?

It's impossible to say whether it will lead to a net job loss or gain because it will absolutely depend on how the value created by machines, algorithms, is or is not redistributed. Marx's point, to stay with Marx, is that the value created from capital left the system because capital owners could not spend it. They stored wealth and that wealth left the system, so you had the immiseration of the working class, and eventually the bourgeoisie middle class. If that happened again, then what you would find is that there would be not enough money for poorer people to spend to support jobs in the wider creative economy, or leisure economy, or whatever economy might survive automation. 

Exactly. Now, if there's a tax system, or some other mechanism that redistributes that value more widely, then people will spend it, and there will be new jobs created. So, it is absolutely up to policy whether there's a net loss or gain. My concern is much more with the spatial sorting of that employment because every industrial revolution, up to number four or five, whatever we're on, has involved moving jobs around between places, and across space. Look at what happened in the industrial revolution: people moving to the cities in massive numbers, countryside wages collapsed. So, that happened most recently to us in the mid 1980s, through to the early 1990s, and has not really been solved in the Valleys and many other parts of Wales. And the issue, of course, with Future Advocacy's excellent report is that, parliamentary scale, you don't see these workplace issues, so the jobs that are lost in the middle of Cardiff as Admiral maybe automates are actually jobs that are lost in Pontypridd where I live and further north, in the—

What I want to do is, I'll come to Catherine for a few opening comments, but then what I'll perhaps do is try and group our discussion a little bit, and we'll perhaps have a look at opportunities that exist. If we can explore that area first, and then go on to some other areas. Catherine, have you got any opening comments with regard to my first question? Or, if I change it a bit: should automation be embraced or feared?

I think the way that we've been looking at it is that automation and the world of technology needs to be led by people, and there are human values we need to kind of embed in the implementation of automation. So, it's up to us how we decide it gets used. I think one of the key things that we've noted is that there are—. We've talked a lot already about the skills gap that exists. You've talked about migration that happens as a result. You've talked about the role of transition between different sectors. And these are things that we are absolutely noticing amongst the companies within out membership. 

In the UK today—. Well, by 2035, there are forecast job losses that automation could displace around 10 million people, meaning that 35 per cent of the UK's workforce could be displaced and seeking new jobs. And that's evident in the challenges I hear you facing. But there's also a huge skills gap, and that skills gap is not just digital skills; it's skills across the board, with 72 per cent of our large companies, and 49 per cent of SMEs suffering technology skills gaps. And that doesn't necessarily mean they are digital businesses, because all businesses need to be digital in some capacity in order to survive in this current climate. 

What we have been working on with our companies, and in this report, is really trying to understand how we anticipate some of those unintended consequences of the technical and digital revolution. And based on the conversations around automation, skills is absolutely huge, but how do we embrace that changing nature of work in its entirety? We have multigenerational workforces now, from all ages, so it's not just about employing young people with digital skills, it's how we retain and retrain our older workers, so that they feel they have just as much right to be in the workforce, so that we're not putting more pressure on our social services. It's how we anticipate that complementary role that automation and humans have. So, how do we support communities to manage that transition? I guess it's asking the businesses that we work with what the workforce planning looks like. It's not just necessarily—[Inaudible.]—what are their labour costs, or their needs? But, actually, most—I think it's 65 per cent—of primary school children will end up in jobs that don't yet exist today. So, how do we start thinking about—? And the job I do now definitely didn't exist when I was in primary school. So, how do we start encouraging businesses to think about that kind of longer term planning and engagement with their communities?

And I think the other final piece is: what is the role of that employer to have some responsibility? I think somebody just mentioned this now—and forgive me for not knowing everybody's names—but what do we do if we displace all of the workers from Ford, for example, but then who's going to buy the cars? We're also talking a lot around the gig economy, and how we start thinking about different ways of working. And we're not saying the gig economy is right or wrong, but it could be that different generations have a lot more—a different way of working, a portfolio career, and actually how we ensure that people have access to the jobs that they want. I think a lot of that is around the importance of infrastructure—digital infrastructure—as well as the physical jobs that we need people to be able to do.

So, we'll be hearing a lot of very similar challenges across the businesses, and I think that your question about should it be something that we embrace or fear—sorry, I'm probably paraphrasing—I think it's a combination of both; I think it's to be really mindful of what the possible implications of automation and AI are. I am actually, personally, a very big fan of AI, I think it has enormous benefits, but it's definitely the role of the people leading the businesses to think about its role and its implementation. That's where we see the biggest differentiator between losing jobs and potentially us having a more vibrant and thriving economy and society.

09:40

Thank you, Catherine. I've sketched down from the conversation so far four areas, and perhaps if we go through each one, it'll bring a little bit of structure to the next hour. First of all, perhaps if we could touch on opportunities for Wales; then we can look at how we minimise the risks that we talked about; and then if we can perhaps move on to the impact of the jobs in terms of gender and skills; and then finally perhaps if we move on to the views on Welsh Government and UK Government action, or lack of action, and then if there's other—. I'm asking Members and members of the panel: is that a reasonable theme to go along? Any other subject areas?

In terms of embracing or fearing, particularly artificial intelligence, which we referred to, and the capacity for machines to think autonomously, is it The Terminator or The Time Machine, as we move forward?

Who are you asking your question to—anyone in particular? Who wants to answer?

I'm interested to see who bites. The Terminator, in the context of the film theories, where intelligent machines eventually sought the extinction of humankind, or H.G. Wells's The Time Machine, where benevolent machines managed the eventual disappearance of humankind.

I would say neither. But I'd like to précis this, in that the reason that we called for the Phil Brown review initially, which the Welsh Government is at the moment setting up, is that a lot of the discussions we're having today, we don't know. So, we don't have a plan in Wales; we don't have the evidence; we don't have sectoral strengths and weaknesses; we don't have a geographical breakdown of our strengths and weaknesses; we don't know our national unique selling points; we don't know our national assets; we don't know the impact timelines and technological tipping points; we don't know the preparedness timelines for the lag in terms of the Government response; we don't know how the public services in Wales should respond to 4.0; we don't have the governance in place and we don't have a coherent network to capitalise on the existing start-up culture.

At the same time, China, last year, launched a 10-year £150 billion AI strategy and the UK Government has already done several reviews: the Dame Wendy Hall review, the 'Made Smarter Review 2017' and the International Labour Organization's future skills review. So, we are playing catch-up. So, what I would say is, speak to Phil Brown and look at what he's doing and the evidence he's gathering, and the outcome of his review, before we decide on a strategy.

09:45

And perhaps we can explore that a little bit more in the last section. Matt, I can see you—if you can just be brief, because I want to try and come on to these block subject areas. Matt.

Absolutely, yes. So, on this issue of The Terminator versus The Time Machine, we would caution that it's unhelpful, really, to characterise artificial intelligence in these terms. The first thing to remember, of course, is that that type of fully autonomous artificial intelligence, or artificial general intelligence, doesn't exist. Although we have made great strides forward in AI, with newer techniques, such as machine learning and deep learning, that can exceed all predictions and beat human expert players of Go—a game where, in 2016, it was predicted that they wouldn't be able to do it for 10 years, and then DeepMind's AlphaGo managed to do that in 2016 itself—that machine will never be able to make a cup of tea, so there's no transferability of very narrow intelligence. But that doesn't mean that we shouldn't take these risks seriously, and there's enough going on with the different forms of artificial narrow intelligence to be talking about, without worrying about AGI.

The second quick point about this characterisation of Terminator is that, if every headline in the newspapers around AI and automation features a picture of the Terminator next to it, then people will feel, 'This is too far removed for me to worry about—this is not going to affect me.' There is some evidence of this. As part of our report in October, we did a YouGov poll, where a representative sample of the UK adult population of 2,000 people were asked, 'How worried are you that AI is going to take your job?' and 'How worried are you that AI will take jobs in your local area?' The answers to those questions were that 7 per cent were worried that AI was going to take their job and 28 per cent were worried that AI was going to take jobs in their local area. So, there are two small points there: firstly, those numbers are smaller than the predictions of how many jobs could be lost and, secondly, people seem to think, 'It might happen, but it won't happen to me', and that's, I think, how we need to take the conversation forward.

Thank you, Matt. I want to start working through these subject areas, so if we spend 10 or 15 minutes on the opportunities for Wales—if we can perhaps spend a bit of time on that. David, do you want to perhaps lead some discussion on that?

Yes, I think it's interesting to flip the coin now and get away from the pessimistic tenor of the last set of questions and look at the opportunities for Wales that might come out of this AI. Of course, this discussion's been going on ever since the Tolpuddle Martyrs and we've gone through automation, where, of course, it was said that automation would free up time for people rather than anything else, and we'd all be on 20 hours a week. That doesn't seem to have happened in any way. So, let's look at the opportunities for Wales, but also look at it in that context: is it going to mean a freeing up of time for people or is it simply going to be that we expand the economy?

So, what are the opportunities for Wales? Who wants to lead some discussion? Leighton.

On the industrial strategy challenge funds, we should be all over that like a rash. The industrial strategy at a UK level has various challenge funds—there is £4 billion, I believe, available. The AI industrial strategy challenge fund was announced a few weeks ago. The Scottish Government have been far quicker off the mark than the Welsh Government have in responding to these. There's much more money at stake than the Welsh Government could—. So, we should be looking at which businesses should be applying to these pots of money and how they can work with universities to ensure that they deliver. For example, 5.0 test beds, there's one in Newport, there may be one in Cardiff; 5G is a prerequisite for automated vehicles. We could be using these test beds as a way of starting an automotive test bed centre in any part of those two areas. And, also, just recognising the assets we already have, we've got places like TramShed, we've got places like Indycube, we've got PyDiff—you know, there is a lot of stuff already happening in Wales that we don't know. So, for example, the head of AI for Europe for British Gas is based behind Cardiff train station. The lead official on ethical AI who is advising the UK Government lives in Mumbles. We're not talking to these people in a strategic fashion, and that's the first thing we could do, which is to ask them what we can do to help them.

09:50

It's a very helpful way of thinking about the risks, but also, in this case, the opportunities to link up different sectors. As I mentioned, we've just done a report looking at the opportunities and the risks for AI in healthcare. So, if I could just speak for a couple of minutes about the great opportunities there, for Wales as for the rest of the UK, there are big opportunities around AI in healthcare, in that healthcare is becoming much, much more data driven. So, people, doctors, clinicians, healthcare practitioners are undertaking and ordering more investigations, greater amounts of imaging data, data from blood tests, and these huge reams of data have got to require automotive technology such as artificial intelligence in order to extract meaning from them, to extract value from them. So, we could certainly have technologies that, for example, improve diagnostics.

In our report, we identified five use cases for AI in healthcare, and apart from this diagnostic element, which we call clinical pathways, there could also be opportunities in process optimisation—so, that's back-end processes around procurement, around staff scheduling. Also in pre-clinical pathways—so, the impact of AI on improving pre-clinical research. So, that's genetic research, genomic research—another area where huge amounts of data are being created. Also in patient-facing applications, and that is where the AI technologies deliver directly to the consumer or to the patient. A really good example of this is currently live in Alder Hey Hospital, for example, where a chatbot is being used to help children and their relatives to ask questions about their hospital stay, such as, 'Where is this particular ward', or, 'What will a blood test feel like?' That gives people the opportunity to ask questions that might otherwise not be asked.

And lastly, population-level applications. So, again, large amounts of data being generated around the spread of epidemics, the spread of infectious diseases, and AI gives us an incredible opportunity to really understand how, for example, infectious diseases such as malaria, Ebola are spread, and therefore be able to predict and institute public health prevention strategies before these diseases hit. So, there are great opportunities in this particular sector.

The other quick thing to mention—

Okay. Can I just bring in Joyce Watson, because some of what you've been saying Joyce wants to ask a question on? Joyce.

I'm particularly interested in the sectoral approach. You mentioned the healthcare opportunities and large-scale data. My next question from that is: who will own the data, and how will that data be used? Because it throws up the whole ethical question, for me, about any life insurance, any possibility that you'll be struck off before you start. All of those things are social outcomes impacting on real people. So, have you looked at that?

Matt, can you hear? What I was also going to add is: in terms of the technology that you talked about, is the technology already existing and not being used, or is there technology already being used in other parts of the world that's not being used here, as well? If you can perhaps comment on that and on what Joyce has said—.

Of course, yes. So, I'll just answer the question about the technology. In early pilot and trial stages, there are good examples of each of the five use cases I mentioned in use. In the NHS in the UK as a whole, there are very few examples of actual, real-world use of these technologies. So, at the moment, there's a bit more hype and hope than actually happening technologies, but the potential is certainly there. Within, I would say, the next couple of years, these technologies have the potential to come on stream. So, that's in terms of what's happening.

To the question around ethics, the actual focus of our report was very much the ethical, social and political challenges of AI in healthcare. What we did was identify 10 such open questions. In fact, our report was done in conjunction with the Wellcome Trust who, when we launched it on 30 April, announced £1 million in funding to start to address the 10 ethical, social and political challenges we identified. I won't go through all of them, but certainly one of the 10 areas, standing in and of itself, is issues around data. These questions are open. Who owns the data? What's the best way to treat the data? We need to remember, of course, that a lot of the expertise and money for the development of these technologies exists in the private sector, and in the UK we have a public healthcare system where the data is, for example, in the NHS's hands.

Another question we had is: what is the best way to regulate or oversee those public-private partnerships? One interesting bit of information we had when we interviewed patients, members of the public, their relatives and we did a round-table discussion with interested patients and members of the public, was that one take-away that I took from that is that patients are very much, I think, ahead of policy makers here. The strong message we got was that, as long as the benefits are clearly outlined of data sharing, and as long as it's clear that people will not come to any harm, for example, by unauthorised sharing of their data with health insurance companies, then people are very happy to share their healthcare data, as long as they know it's going to help either them or people with similar conditions or their children in the future.

We certainly need to be having a much more nuanced discussion. There have been bad examples of how this has been carried out in the past that we should learn from. With that, I have in mind the Care.Data issue in 2013 and the issue between the Royal Free Hospital and DeepMind more recently, which the Information Commissioner's Office reported on. So, there have been bad examples of doing this in the past, which we need to learn from.

09:55

Thank you, Matt. I'm just a bit conscious I can see our conversation lasting until 8 o'clock this evening because there's so much detail here. We've got 45 minutes left in this session so I'll just perhaps say that we're only going to be able to touch the headlines, I think. But keep in mind as well when you're commenting that what we want as a committee is to make recommendations to Government, so we want your views on what you think the Welsh Government should be doing. Catherine, you indicated you want to come in.

It will be very short, bearing in mind we have to move on. But, just to echo some of Matthews's points, we worked with a start-up a few years ago called Mastodon C, which was working with the Open Data Institute. They had used data and analytics to identify that, just through identifying where branded statins were used in particular healthcare trusts versus generic statins, if they could use exactly the same type of generic statins across the board, a saving of £27 million per month could have been made on one type of drug. So, there is an enormous amount of information out there that could be used to help in a variety of ways.

I want to make sure that, when we're thinking about data, there is a lot that needs to be done in terms of protection, as Matthew said as well. We interviewed the CEO of Bupa this week, and I work very closely with Aviva, so we do talk a lot about the role of especially insurance companies, whether it's health or whether it's general insurance or life cover and how they use the information for the greater purpose and how they start thinking about its implementation in understanding trends and what we can do to prevent and support systemic issues rather than targeting specific customers and ensuring that, just because you're using anonymised and aggregated data to track and identify trends, we're not punishing the individuals or we're not protecting them as well as we should do. I know you want to go back into the main questions.

Thank you, Catherine, yes. Can I ask Members if they've got questions around opportunities that exist for Wales at all? Mark.

In terms of new possible jobs for human beings, what opportunities do you think might arise? I know there have been some more optimistic forecasts about the potential new jobs that would be created for people.

10:00

Can I just hold that until our third section? In our third section we've got impact on jobs, skills, and perhaps we will hold that until then. Just looking for any other opportunities that exist for Wales—perhaps looking to Calvin.

There are two things. Certainly, in terms of the here-and-now specifics, there are one or two areas where we can see, perhaps, something of a small advantage. Cyber Wales is not spoken about very much. That's a grouping of a large number of smaller companies looking at cyber security, cyber defence, with the Wales network that I think could bear more—I'll say, bear more policy scrutiny; I'll kill it now, but you know what I mean. The National Software Academy in Newport and various bits and pieces spinning off from that. There are areas where we are doing stuff on the ground, albeit, I think, fairly small and certainly not engaging with the vast majority of Welsh workers.

I think, more critically, what hasn't come out so far this morning very strongly is the policy framework under which this stuff goes forward, or doesn't. I was talking to colleagues in the university last week about the possibility of setting up a new network looking at policy innovation in small open regions and countries, because Wales has space to do stuff you don't have in the west midlands or even in, maybe, the Manchester city region yet. We've shown that so far with the future generations Act. It's interesting that, of those 350 objectives, coming back from the 44 agencies under the Well-being of Future Generations (Wales) Act 2015, virtually none of them are saying, 'Respond to automation'. They're all very here and now, so there's an issue there in terms of how the public sector are responding. And the opportunity is to explicitly fashion policy and explicitly fashion—both apply certainly for UK-level funds—regional funds in order to incentivise behaviours that develop resilience insofar as we know how we do that, which I certainly agree with Leighton that we don't really know so far.

I was talking to colleagues a couple a months ago in Brecon Beacons national park. I was thinking that nobody, as far as I know, anywhere in the world, is envisaging, let alone seeking to build, a small rural community that is connected, progressive, inclusive, e-connected, has employment, has social opportunities for people, creative opportunities. What does that place look like? Now, that's the sort of thinking that I think the Welsh Government hasn't done. What's actually going to be viable in 2050 and what's actually going to be properly productive, not just narrowly productive, in 2050? And then, working back from those sorts of visions to the sort of policies that might start to put that in place—

Yes, sure.

That's a very interesting idea. Would you envisage that as a kind of test bed, an eco-town of the future? Or are you thinking about—retrofitting isn't the right term in this—future-fitting an existing settlement?

I think, certainly, one would probably do this at city region level. South-west Wales, for example, albeit that city region's gone in a particular direction that maybe doesn't help this. But you would probably do it—. It would be mediated by whatever structures are in place that would allow certain investments to happen. But I think that that sort of—. Now, I'm absolutely aware that, in Wales, we do lots of pilots. So, we step in and out, we have policy interventions, stuff happening, and we go away again and we evaluate it, and badly usually. So, we would be talking about a period of something like a 50-year city deal, and actually choosing a place, maybe via a competition or based on a set of pre-existing characteristics, where we think we could look at what a—I'll say futureproof. What sort of education facilities will that place have? What sorts of employment sites owned by what sorts of people? And what sort of cyber security? What sorts of 5G wire and other connectivity would it need? Having those sorts of thoughts, with a vision in mind, even if the vision turns out very different to what you expected, it would never have been wasted, would it?

Just picking up on Leighton's point earlier, in an urban context as well is there an opportunity to build an urban innovation test bed? Smart city is possibly an old-fashioned term now. In Tokyo, Dubai, Singapore, they're talking about creating robot cities or fully automated cities. Is there some advantage in actually choosing a designated area, an urban area, and saying, 'Let's actually use that as a full-spectrum test bed'?

10:05

Yes, but—just to finish the point—only if you do it properly. There's no point doing this in—I'll say 'half-arsed'; I don't want to say it but I will say it—a half-arsed fashion, with a little bit of money from other pots, without actually changing the physical fabric as well as the kind of connectivity fabric of the place, without changing, maybe, even things like tax laws. Could you envisage a place where you had land tax, so that people would use land differently? That's the level of change that you would need to try out to see if you could actually make a difference.

Similar to the kind of special economic zones you see in other jurisdictions.

Yes, but under the kind of frame of the future generations Act, which says, 'These are the things we intend to do, and if we're going to do them, these are the sorts of things we need to try.'

Just on the tax point, we won't agree on the tax point, in the sense that tax needs to be a last resort, not a first response. I think if we revert to tax levers, as others have said, in a different place, around taxing robots—. Could you imagine, 10 years ago, taxing mobile phones? I think that's a result of a crisis-led engagement. If we don't do anything in the next few years, the debate will end up going to taxing things, and that's an example of a failure to address this, rather than embrace this. As a very successful business said the other day, there's no reason why the NHS should be short of money; it's sitting on a goldmine of data. Now, it's not easy to access, it's gnarly, but it has the potential—

It's on paper. 

It's on paper. But we've got to do this sooner or later, and if we don't use the data—. The UK Government are looking at data trusts, so you provide a platform—. Because accessing AI is easy—lots of universities have access to AI—but your product needs to be tested with reliable data. So, hopefully, the Welsh Government should have data asset analysis and actually see how it can be open. Cambridge Analytica has shown that we need to use this in a responsible way, but not doing anything with it is criminal. 

I'm going to come to Mark in a moment, and then can we think about moving on to the risks that exist due to automation—how we minimise the risks? If perhaps the discussion can move that way. Mark, you wanted to come in.

Yes. You've just described a future focused on things and systems, but how do we move from needs-based approaches to asset-based approaches, where those assets are the most important building block—the human beings themselves—unlocking their strengths and ensuring that we're delivering these broader potential systems or infrastructure with them, rather than simply doing it to them, which will simply perpetuate the same social problems that we were seeking to address in the first place?

Well, I think you do it through a programme of dissensus, if you like, by actually—. I mean, the idea of a—I won't say 'eco-town'—but some sort of futureproofed rural community would arise from what the people who currently live in that community wanted. And the programme would be—I won't say 'a programme of education'—a programme of engagement, so that everybody has as full as possible knowledge about what the upcoming challenges were. Then you base your interventions on what people say they want. It's ludicrous to me that my son comes home from school and says, 'I touch a computer once every two weeks' because the teachers never say to him, 'What do you want to do in school today?' Whether it's schoolkids, whether it's people who live in the middle of Snowdonia being told what goes on their hills by Natural Resources Wales in Cardiff, we've got this way in Wales of this kind of top-down, blanket, one-size-fits-all approach to policy, which in this case won't work. Because the stuff that's going to work in south-east Wales in this area—. South-east Wales may end up with a fantastic, the most secure city-wide—

Thanks. I'll come to you in moment, Matt. Mark, did you want to come back, or not? Are you happy with that answer?

Well, I just wanted to try and move it on, so I'm not sure if you want to bring Matt in first.

Matt, if we want to try and move the discussion on to minimising the risks that exist, is your comment around that, or is it around the opportunities that exist?

I was actually going to move on to minimising the risk, and picking up on Calvin's last point.

In that case, then, could I ask Lee to ask this question? Then perhaps you can come in on Lee's question.

This is a fascinating session, but it's so vast, isn't it? We could have sessions on the health and the education elements by themselves. We have a session coming up later this morning on the agricultural elements, as a case study, if you like.

Can I just go back to the evidence we've received and try to bring in Catherine Phillips, who mentioned—? I'm just touching on Leighton's point about taxation as the last resort, which I think is a powerful one. In the Business in the Community report, Catherine, you say that AI could increase labour productivity in the UK by 25 per cent by 2035, so clearly there are going to be benefits to business overall from this. Without straying back into Marxism, in terms of how we try and capture the benefits of this so that it's easily spread and it doesn't turn people against technology, it sees it as an overall benefit to society, bearing in mind that the Welsh Government has limited powers or economic levers, how do you think the business community feels that the benefits can be spread and shared?

10:10

I'll come to you, Matt, after I've come to Catherine, if you want to comment on that as well. Catherine.

The question is how do I feel the benefits could be spread—

The productivity you say is going to be increased—a 25 per cent increase in productivity. That's profit. That's a good thing. How do you make sure that it's not just held in the hands of a few so that people don't turn against AI as a threat to them?

I think it's about the comment I made earlier, which was about how you ensure that the human values are instilled into the purpose of AI. So, what is AI there to do in the first place? Is it just to create more productivity? Actually, is it to strip a company so there are greater efficiencies, but actually it doesn't make a more effective business? It comes backs to that point earlier that if there aren't people employed in order to buy the products, does that mean that the productivity has been worth it in the first place?

We are a business-led organisation and the way that we want to work is with the companies in our network, to ensure that they are creating mutual value for people and for the productivity, for the use of AI—so, how do the roles of technology and humans complement each other? How do we think about the role of technology in preventing health and safety issues? What does that mean for when we're building the next HS2, for example, or Crossrail? There was an example of Accenture and Airbus using HoloLens glasses to try and minimise any potential accidents when they are creating the next air fleet. So, there are lots of ways that we can start thinking about productivity, rather than it being solely to replace people in their roles.

But it does come back to what are the drivers for it in the first place. I think that Matt touched on the point around media and the way that it's perceived. Actually, a very useful point of thinking is that, quite often now, AI will offer one particular solution, but general AI doesn't yet exist in the way that we think it will exist, which means that we don't have the robots that will be taking over the world. So, if we can find specific solutions that help us tackle key issues in our businesses and in our societies, then I think that helps. So, whether it's the NHS issue with the data mining so that we can understand how there are potential huge cost savings—that's one really great example. But it is driven by our business leaders—

I was going to say that the key thing is not presuming that it's led by anybody other than us—

I was trying to interrupt you, because I'm trying to focus this conversation. It seems to me that one of the things we could do usefully is to identify particular problems that we as a country, as a society, want to solve and try and martial AI around that and try and martial the different agencies to try and fix particular problems. Otherwise, this seems to me a very diffuse debate. And because it is such a rapidly changing environment, is that a way that we can tackle this—to turn this on its head, and rather than see what technology evolves and how we then try and respond to it, as a country, as a Government, to say, 'We have these three problems'—or five problems, whatever—'come forward with solutions'? Would that be a way forward? I'm struggling to understand how we as policy makers try and shape this to benefit our communities.

10:15

When we conducted our research that led to 'A Brave New World?' we interviewed over 50 CEOs of businesses across the UK and they informed our decision making on what were the key issues. So, 'A Brave New World?' set out the risks and the opportunities and then we identified the four things we wanted to tackle and very practical business actions that needed to be implemented as a result. It was data, the changing nature of work, how we create inclusive product and supply chains, and what are we doing to tackle transparency and the environmental issues. So, we, as a network of businesses—and this is publicly available information—have the research available that, alongside the digital strategy that the UK Government launched last year, sets out connectivity and skills, the wider economy, cyberspace, digital government and data. These are all overlapping, and it's about picking the ones that are most relevant for Wales, for your economy, for your societal issues or challenges, and how you identify where there are very practical and tangible actions, and working out who are the key individuals. Many of them, I'm guessing, are around the room today. You've got to prioritise, I guess, which are the biggest and the most important for you.

Thank you, Catherine. I was going to say—Mark, did you want to come in at this point? I think, Leighton, you were just indicating as well. Did you have anything to add, Matt? 

Yes please, just to try and bring a lot of these points together. So, on Lee's point, I agree; I think a sector-specific approach is rational here for two reasons. Firstly, it's because of the nature of the technology. We're dealing with artificial narrow intelligence. It's very task oriented, there are specific tasks in different sectors, and there are technologies built for different sectors. Secondly, in terms of the regulation and ethics, dealing with healthcare data is very different from dealing with transport data, which is very different from dealing with agricultural data. So, a sector-specific approach is the right way forward, and the UK Government, particularly with Dame Wendy Hall's review in terms of the creation of data trusts, which Leighton mentioned, I think is a good step forward in this direction. So, that's on Lee's point.

Then, on the other points in terms of the risks and how to minimise them, in our report we make six recommendations. I will just, at this point, pick up on the first one. One of our main concerns is not really about job creation and job loss and whether it's going to be net loss or not. Economists disagree very much on the total jobs created and jobs lost, but it's about the potential for AI and automation to ramp up and to accelerate inequalities. That's why I very much agree with Calvin's point that one-size-fits-all solutions are not going to work. We need to really better understand who is most likely to be impacted, and where they're more likely to be impacted, in order to target solutions there. So, our first recommendation is for governments to commission research in order to understand this better.

We're very pleased to hear that the Welsh Government is commissioning the digital innovation review that is looking at this. When I met Phil Brown at an event that Leighton organised a few weeks ago, I said the type of research that needs to be done here is very much community focused, and this will have two advantages. Firstly, if you go to different communities and you get together local business leaders, local unions, the local AM and the local MP, local employees, put them together in the same room and look at the local challenges and how you're going to solve them, that is one way to get local solutions, but also to start that conversation on how these technologies are going to impact those local people. So, a local approach is what we're calling for, and we hope that the digital innovation review will take that forward.

I'll bring Leighton in in a moment, and then after that, Mark, I'm happy to come to you for you to ask your question around jobs and skills. If we move the discussion in that direction, then I'll come to Joyce and Vikki on that same topic as well later.

Three points in 30 seconds, hopefully.

The economic contract—the Welsh Government are developing that. Why not add an element—it's Government money, the Welsh Government decides what to do with it—around ethical employment practices? Why not have an element in there around AI and automation and the impact on those businesses?

The second thing is that, ideally, this should take place at a supranational level, so we avoid businesses going from place to place and competing against each other, and if not, at a UK level. That's why maintaining the UK internal market is so important, because we want to avoid, I would suggest, at any cost an Amazon situation, which is happening right now in America, where they are playing state against state for the best amount of government money for their second Amazon headquarters. I think that's probably not where we want to end up. And the final thing is around Germany: they did this first, automation and robotisation. There is a model there where they didn't really lose that many jobs. They understood that there were highly skilled people in their workplace and they developed it in a way that repurposed those individuals and had a situation where there were cobots, so you had a worker and the robotics systems working in unison. So, that's a model we should look at, but, again, early action is critical.

10:20

Sorry, just on your first point of the economic contract, about putting AI in: isn't the Government saying it's going to do that?

They may be saying it to you.

Well, I'd understood, as part of their economic contract, that preparing for automation is one of the criteria.

Well, it's not in the one I've seen, but it could be in the one that—

It's in the economic strategy, isn't it? It's in the action plan.

So, the economic contract is a layer down— 

—and it's going to be unveiled shortly, I believe—

—and there are certain criteria. And it's—. I don't want to be speaking for the Welsh Government, but it's pick and mix in some degree, and it's—. The final version, as I understand it, will be subject to the fair work commission's outcome. So, that fair work commission has to do its work, so I would say those who'd support that should be submitting evidence to the fair work commission.

So, rather than pick and mix, we should say this should be a must.

Pick and mix in certain elements, but, you know, if we wanted to say, 'Look, certain standards of automation when you're using Government money', then that would be something that you could consider.

Okay. So, if we spend, perhaps, 10 minutes on job skills, and then we finish with 10 or 15 minutes on what the UK and Welsh Government are doing or not doing. But we do want to have some recommendations as a committee to take to Government, so perhaps that last 15 minutes could be based around that. Mark.

Yes. I think you may recall my question. We've heard a lot about the threat to jobs and the possible social and philosophical consequences that that might bring, but what opportunities for new employment might this bring for people either currently employed or entering future labour markets?

Can I just—? Before anyone else answers as well, do you mind—if I mention to all four of you—if we're not quite hitting the point, we interrupt you? Is that okay? Yes. Thank you. I'm very grateful, because we're really short of time. Who wants to address that point? Calvin.

Yes. As Leighton said to start, we don't know. We have no idea what new occupations, tasks, jobs, even industry sectors will be created as a result of—I say automation, but certainly artificial intelligence, automation, robotisation, algorithms, you know, all the industry 4.0 stuff. So, we have no idea—. If little Johnny currently gets pointed towards being a bricklayer, we have no idea what it will be in 20 years' time. And that means that we have to—. You made a really good point earlier about moving your asset-based approach. So, at the moment, it seems to me that if we trained, educated, helped our children and our adults learn how to understand information and develop intelligence in complex scenarios, you know, address wicked problems where you've got lots of different social, technological, economic contexts, and to respond to those with measured and practical kind of actions, then that wouldn't be a bad thing, and I don't think that's happening from what I've seen of the work I've done recently for Welsh Government with schools. This idea that we need to teach with a view to the complexity of real-world situations is something we are desperately trying to move towards in my university and my school, and I think the Donaldson curriculum—curriculum for life, whatever it's now called—coming in in 2022 is the last opportunity to do that in Welsh schools and to address the chronic underskilled, underqualified nature, in most regards, of our population, because whatever happens, whatever jobs are created, they won't be taken by people who are insufficiently engaged, and the way you get engagement is by education, largely. So, 'education, education, education' is probably the only way that Tony Blair's got still standing, I guess.

10:25

Yes. Certainly, again, it's very hard when people don't agree on what new jobs will be created, but there are two general principles that might be helpful to keep in mind: firstly is that technology is advancing at an accelerating pace—and this is all technologies, not only artificial intelligence and automation. So, the real skill and the policy implication is the need for lifelong learning, the capacity to retrain. We are already, and we need to definitely move away from this model where we get all our education in childhood up to the age of 16 or 18 or 22 and then bank that education for a lifelong career. That's not going to work. It doesn't work in many cases now, and it's certainly not going to work with things changing so quickly in the future. So, we need to embed lifelong learning, and systems for lifelong learning, in education systems. So, that's No. 1. 

The second thing is to keep in mind the types of skills that are more likely to be resistant to automation in the longer term. And, by that, I mean creativity—not only the arts, but being able to think on one's feet—but also interpersonal communication skills. And we think that there's a real opportunity to value professions that currently are undervalued, such as social care, care work. We have an ageing population, we have a dearth of care workers in all parts of the UK and across the developed world, in fact. So, there's a real opportunity to shift focus towards more caring professions and to value these jobs properly so people see them as a useful and as a good career option going forward, because these types of jobs are less likely to be automatable. 

I've seen some media coverage giving positive forecasts and actually identifying potential new work opportunities in the consequence of the automation revolution. But are any of you aware of any of that research, or is it purely anecdotal and journalists putting sticking plasters of different articles together?

There's lots of research. It's whether you believe it or not is the problem. So, Nesta has done a range of work on this about jobs in 2030, which range from the sublime to the ridiculous. So, it's whether you believe that or not. So, there are things like data detectives, green construction, AI-assisted healthcare, cyber city analyst, genomic portflolio director—all of these things. The problem is we don't know whether they're going to exist, and we don't know whether they're going to be here. So, we need to understand what our unique selling points are in terms of our skills base in order to set a plan to prepare ourselves for potential jobs.

The one thing I would say is that there are potential opportunities. Swansea University is the first in the UK to do an LLM in legal tech. So, what we haven't talked about is that the paralegal situation in the UK, and particularly in Cardiff, is very strong, and they are at risk of automation. There's one firm in London at the moment trialing a bit of tech that can look at non-disclosure agreements in 150 documents, I think, in 20 seconds; it takes a lawyer 90 minutes. So, what happens if that gets rolled out to our legal services in Wales? We need to be preparing these people to retrain in legal tech areas. Ten per cent waste in the public sector. In the NHS in England, we spend £15 billion a year on admin. What could we be doing if we had a more clever approach to that and recirculated that funding to more productive, front-line services? And the other thing is around Go Compare. I was with them yesterday. They're spending £40,000 per student to train up new coders, so the money is there, but, unfortunately, they're using an English university, because Welsh universities are used to the supply-led model not the demand-led model. So, we're falling behind there. So, businesses will take advantage, and they will try and get ahead, but it's whether we're doing enough here. 

Okay. I'm looking at—. Is there any—? Before I move on to questions around gender and age, any other questions on what's been said before I come to that point? No. 

On skills. Go for skills, and then I'll come to Vikki on gender. 

Right. I just want to stick with what Leighton's just said about coding. The Minister's clearly got her eye on us training people, but we've got to train the trainers. And I don't know whether you're able to answer this, but, if you are, where are we in terms of having sufficient people in place who can actually teach coding?

10:30

We don't know. So, the data that's available—we looked at this the other day—is at EU level; it's broken down at UK level. So, hopefully, this Phil Brown review will actually access that dataset and try and have a better understanding of where we are in Wales. I'd imagine we're not in a strong position because these are highly valued jobs, and they tend to go where they're paid most and they've got better life chances. So, it's kind of: 'What's our offer for these coders to bring them down to Wales?' I think we have a strong offer. Those places—Oxford and Cambridge, Silicon Valley, the Silicon Roundabout—aren't perfect, and there are many things we can do to attract them to the M4.

Training coders buys you a few years, because, arguably, there are already coding algorithms, and one of the dangers is that we take our—. You know, what we don't want to do is take our current silo discipline-led approach in schools or adult education around things like geography and religious education and history and replace it with another bunch of silos. You know, I did BASIC—much good it did me or would do me now. It's about teaching the principles of coding and the idea of how intelligence works without locking students into a particular language, for example. Python's all the rage, or was all the rage two or three years ago; it's moving on now. So, it's understanding what level you want and what sort of competencies, rather than skills, you want to engage people with so that those competencies can be re-oriented when it becomes clear which language in that case will actually be used. That's quite difficult. 

Can I bring Matt in as well? You wanted to come in, Matt.

Yes, and it ties in, actually, with one of the points that's been made around gender, which is that, with coding, I agree it's a short-term, short to medium-term, fix. However, there is stuff to be done around coding, and that is predominantly with getting more women and people from ethnic minorities into STEM subjects and coding—so, science, technology, engineering and mathematics. This is because—. For example, Kate Crawford, who's a professor at New York University has spoken about artificial intelligence's 'white dude' problem, which is that biases that are perpetuated by algorithms are not picked up on or not treated as important if everybody who's developing AI are middle-class white people in Silicon Valley. So, we need much more diversity of coders and of people in STEM so that they can be more representative of society as a whole, and therefore the technological tools they create serve society as a whole.

And, Catherine, you wanted to come in. Catherine. 

Did you want to come in, Catherine?

Sorry, yes I did. Sorry, I had muted myself. Very quickly, I totally agree with what Matthew was saying as well. One of the key points we would recommend is that digital on its own is not necessarily the answer; it's a lens through which you can address some of the other issues. So, whether it's around the subjects that you're learning in schools—so, whether it's geography or learning about farming, potentially, or learning around any other subject, it's so rapidly changing that we would encourage—. The skills that you need to learn are not solely digital skills but they're skills that can be implemented and used across a variety of other subjects, and that's what we think would give, potentially, the edge to future employers that you're not solely working in a digital industry. Not all of the businesses that we work with would identify themselves as having a digital element. So, it's around, I guess, making sure that the variety of skills are included, not just seeing coding as the answer. I think that our more advanced businesses are more progressive than that.

Yes, certainly. It seems to me that perhaps one of the best things we can do to harness the power of automation and to safeguard against the negative effects as well is to gain a greater understanding of exactly who's going to be affected by it. We've talked about different sectors already, and retail is one that I've seen that could be very strongly affected. So, that leads me on to thinking is this an issue that's going to affect females more, because of social care as well. Should we be looking at a certain age of our workforce to prioritise assistance as well?

So, we need just brief comments on that, because we've got to move on, but Leighton I can see you nodding.

So, yes, they will be impacted negatively in the first wave in terms of retail, but also, with care, there's an opportunity to make it more socially valuable, and women are predominantly in the caring space. Also, if you look at the new skills that are most valued—empathy, creativity, listening and learning—I'm not saying men don't have them, but they are more common in women. So, they are well positioned, but we need a strategy to support them.

10:35

So, yes, there are a couple of good reports around this. I'll point you towards PricewaterhouseCoopers's report just from March this year, I believe, which divided automation into three phases and suggested that women are more likely to be impacted in the initial stages of automation, and that is because the types of roles that are more automatable in the initial phases are more gendered towards women, and those are more in data input, clerks in hotels and retail and things like that. But in later phases—and what we're talking about is 2025 to 2030 approximately—men maybe are more likely to be impacted. That's to do with more driving roles, transport roles, autonomous vehicles coming on stream. So, we need to factor in both time and gender, but I completely agree with you—the initial point is that more research is needed to really understand who is most likely to be impacted.

I think the age piece is really relevant. For people over the age of 50, it is a lot harder for them to even get through the door of an interview, let alone thinking about when there are redundancies being made. So, between the ages of 50 and 70, with our ageing population, yes, that retraining and kind of retention piece beyond even recruitment is incredibly important. We're having these multigenerational workforces and we're talking to companies about how they maybe use some of the insights and intelligence they have, and how they're training up the younger generations on digital skills and actually thinking about not just porting over, but really identifying what the potential issues are for their older workforce and how they take that learning, so that we don't have kind of a more paralysed and divided workforce as a result.

Thank you, Catherine. Any further questions, Vikki? I'm just keen that we keep them short—short questions and focused answers—because we've got to move on to the last section. I've got Adam and Mark waiting. Adam.

Okay. A short answer to this would be appreciated: should we be warning businesses, for example, about the risks of over-automation? I don't know if you saw the comments recently by Elon Musk about the problems of the Gigafactory. I never thought I'd see it, but he said that humans are underrated. We put in too many machines, and they're having to bring the humans back. So, should part of the role of Government be saying, 'Look, automation isn't for everyone'? 

We could have a longer answer on that because that does stray into our last section, so I think that deserves a longer answer.

I think Government—[Inaudible.]—anywhere near enough to make that judgment call for individual businesses—maybe as an overarching 'Be careful'. But again, with a particularly Welsh kind of hat on, the people making the decisions aren't in Wales and are not listening to you.

As a general point, though, maybe this is bringing the au contrarian out in me, but the fact that one of the biggest apostles of innovation of our age is saying, 'Actually, I think I got this wrong. We're having to bring the humans back'—that's pretty significant, isn't it?

He's an outlier, I think, because he wanted to completely automate his factory. I don't think a lot of businesses are there yet.

Again, in Wales, the problem is that the vast majority, particularly of SMEs, are completely unengaged with the automation agenda. I completely take the general point. I think, actually, on the ground in Wales, when we talk to businesses, where there's no superfast broadband, tourism or various other things that we're doing, this is just not on anybody's radar, and it's much the same in the public sector.

Okay, to flip it, then—doesn't his insight actually, going back to the earlier discussion, shine a light on—actually, automation highlights the things that humans, and only humans, are really good at, which is complex, creative, collaborative problem-solving?

He'll hate me saying this, but it's worth getting Tegid Roberts in to talk to you about this. Tegid was, I think it's fair to say, fairly instrumental in getting the Raspberry Pi made in Bridgend, and he says that there are hundreds of people now working in Sony Bridgend making a Raspberry Pi and none of them touch a Raspberry Pi. It goes on to the the conveyor belt at the start, as a bare motherboard and comes off the end as a computer, and nobody touches it in the meantime—they're all doing other things. So, it's about understanding this kind of labour-augmenting automation in combination with the moving of people away from 'manual tasks', repetitive tasks, into things that are more creative.

10:40

And most firms we speak to about automation are doing it to sustain their presence in Wales, because of the risks of Brexit, risks of access to labour—they're not doing this as a cost-cutting measure, principally.

Yes. Elon Musk is an outlier. [Inaudible.]—all sorts of things: they're talking about how he'd structure a government in Mars as well. So, we can't really take him at his word on everything he thinks about. But there is a general principle we can take, which is that, yes, humans are underrated, and therefore the risk is that we fall for the hype of how much automation can do, and too soon, essentially. So, what the Government should be doing is encouraging a much more nuanced view for businesses. And remember that there's a lot of interest in AI. McKinsey in 2016 estimated that $40 billion was invested in 2016 in AI globally. So, there's a lot riding on it, and one of the risks is that we fall for the hype, move too quickly, without considering what humans are still good at.

I've got Mark waiting. We've got 10 minutes left, and I'm really keen we move on to—I want your views on action that Welsh Government and UK Government have undertaken, or lack of action, and, particularly, what they should be doing. So, Mark, if you can shape your question around that.

Well, you've spoken about the impacts in terms of gender and age. What about disability? Disabled people are already under-represented in the workforce, and often over-reliant on employment in supermarkets—Tesco et cetera—which you indicated, Leighton, are more likely to be taking the short-term hit. So, how can we turn that, in terms of futureproofing, and calling on Welsh Government, into positives? I'm thinking particularly about how AI might help remove the barriers that disabled people face in the workplace, but also about better accessing the highly relevant skill sets that many people, particularly on the autism spectrum, and otherwise, have, which are applicable to this very agenda?

It's difficult. I think there are apps that some people can use. So, people who are short-sighted, there are already apps on the iPhone that will describe what the camera is seeing, to aid people in the workplace. But I wouldn't know enough on that to comment, I'm afraid, but we can get back to you—we can certainly write to you about that.

This is very difficult. In terms of physical disability, I think there are opportunities around the decentralisation of work, and the ability of people to work closer to home, which are—. As I say, we're working on a project looking at some of the opportunities around superfast broadband, and the ability to do work away from traditional employment centres. And I'm very wary about the idea of teleworking, everyone sitting in their bedroom all day—disabled or otherwise; I think it's very dangerous to assume that's what will happen. But, certainly, I think shorter journey times to places that are well connected, which have a wide range of supporting social and other infrastructures, which would enable people to be productive closer to home, is a huge opportunity, particularly in the case of the south-east Wales metro, which we should be thinking about very carefully.

Okay. We're keen to capture from your answers now what Welsh Government should be doing. Lee.

I'm finding this whole session incredibly frustrating, I must say, because it seems that where we need to be going to quickly, and where we're at, there's just such a profound disconnect. I was struck by what Calvin said earlier about his child's experience in school, and not being able to access an iPad, and I wonder if there's a parallel there. Because if you look at the way schools are using digital and IT, there are massive divides. So, you have some schools, and they may not be in the places you would expect them to be—so, for example, Cadoxton, in a poor part of Barry is an outlier of this; it's really very innovative in what it's doing. Other schools—I went to a private school in my constituency recently, where they still had a computer suite, a secondary school, which is frankly from the ark. So, there's a lumpiness in the way that schools in Wales currently are approaching this, which is 10 years behind the curve, and I wonder if that kind of lumpiness we are likely to see in the way these technologies are adapted in practical ways, and the way Government's responding to it, because you've still got the Government talking about—. We've discussed coding could soon be wiped out—sooner than in 10 years—replaced by algorithms themselves, and we can't even cope with coding.

Now, the positive is that the Donaldson review does talk about creativity, and many of the things that Matt Fenech recommended we should be talking about. But just in terms of the pace of change, and the ability of Governments, but ours in particular, to respond to it. I note today there's a job advert out for a senior Welsh Government director for economic strategy that does not mention digital, AI, automation, or any of it—it doesn't mention it. So, I just want to scream, really. But in terms of practical steps and what we can do, what would you—? There's a Phil Brown, I think, review commissioned, and my fear is that we're going to wait until next January before that reports and then nothing happens in the meantime. 

10:45

I'll answer as briefly as I can on this. The problem with the Welsh Government is that it spreads itself like the thinnest butter on the largest piece of bread you can imagine. It tries to do all things, and it does all the things really quite mediocrely. As I get older, I'm increasingly convinced that the only thing that matters is the education competency skills of our people, whether they're workers or otherwise, and their engagement. I think that if I were king of Wales, or a gender-neutral description of the leader of Wales, I would just borrow up to the hilt, and I would spent £1 billion or £2 billion on making the Donaldson curriculum, when it comes in in 2022, as technology resilient and technology focused, and as well funded, as it could possibly be, and I would probably, as part of that, have to be a little more overbearing in terms of school autonomy and headmasters' autonomy. Currently, headmasters have a lot of autonomy, which ends up in the slumpyness that you see. Their priorities might—. Usually, I'm all in favour of subsidiarity, but I think in this case, there may need to be a bit of a drawing together of a little bit more standardisation around the level and engagement of technology, not just in terms of it being an IT lesson, but in how they access information across a wide range of subjects. Donaldson does give the chance to be more interdisciplinary, certainly within humanities and science, technology, engineering and mathematics, and other things. And I think, even if we didn't end up spending that money in a way that made Wales resilient to Industry 4.0, how could we possibly waste money by spending £1 billion on education in Wales to make it more modern? And I think I would just do nothing else.

Can I just, on that—? At Atlantic College—we went to see them the other day—it's not about money, and they're obviously a privileged private school, but they designed and built their own 3-D printer two years ago in Atlantic College. They're looking to have a start-up cluster on site for their school and students to benefit from. It's not a money issue—it's not going to be very costly—it's about leadership—

It's got money, though, hasn't it?

No, it isn't about money, though, because the 3-D printing stuff was donated and the start-up building is going to be an empty shed that they've got—

But without getting into a debate here, they have the human capital—

You know, Cadoxton doesn't have money, so it's about the leadership.

What they have is a network of former students who are willing to donate their time, and what Wales has is the same thing, but we don't use them. So, there are world leaders in this area across the UK and across the world who we never speak to. Why? We could ask them to contribute something back to Wales.

The second thing in terms of what we should be doing is: at the moment, we've got a hole in the bottom of the bucket, so I'm going to mention my favourite hobby-horse, Ambionics in north Wales, which is an incredibly good business—a 3-D printing start-up—which wants to be a social enterprise, not-for-profit, and is leaving for England because the business support structure in the Welsh Government doesn't understand how to deal with 3-D printing businesses. So, I know about that one, but how many others are in bedrooms and in garages across Wales, which are thinking, 'Actually, the business support structure in Wales doesn't talk to me.'

We're running just over time now, so can I ask each of you for your top recommendation for Government? You can have up to three if you're succinct with the other two points, but your top recommendation, or your top three—if you're succinct—recommendations to Government. Who would like to go first? Leighton.

Asset maps—what are the private sector strengths and what are the public sector strengths. Skills gaps—what the rate of our skills base is in Wales, and match those against the jobs most at risk from automation, and put skills pathways in place. And look at the start-up community and also the existing businesses and look at barriers to adoption—4.0 doesn't challenge businesses, it challenges business models. So, if we do that, that would be great.

Perfect. Thank you, Leighton—succinct. Anybody else? Catherine.

I'd echo the point around talking to businesses. So, obviously, we are a business-led network, but there is an awful lot of appetite out there to address some of these huge systemic issues. Rather than relying solely on Government, what is the role that the businesses that are the employers, that are the future educators, can play? We are hopefully going to be launching a huge global network of champions. 'Global' doesn't mean it's restricted, so I would definitely look at who the key influencers are within business. Secondly, I know our priorities that came from the report are aimed at business, but there are a lot in there that are very easy to adapt to thinking about the role that Government can have and can play. So, my two recommendations are probably focused on those two.

10:50

I've already mentioned two other recommendations, which are around taking a community, local approach to further research and having local solutions, and reforming the education system to value and to provide lifelong learning. The other recommendation that I haven't mentioned is around this issue of taxation and redistribution of the increased productivity, and that is to look into alternative income models, such as, for example, universal basic income. At Future Advocacy, we don't think it's the panacea that some people might be promoting it as, but certainly it deserves further exploration. There was a Finnish trial that has just ended, so we should be getting the results from that next year, but we know that the Scottish Government is funding four UBI trials in Scotland, so maybe this is something that the Welsh Government should look at as well.

I don't disagree with anything that's been said before. It's a very embryonic thought in my own head, but I think this idea of pressure-testing what a productive community, rural or otherwise, looks like in 2050 with some serious money is worth consideration. The second thing is that, somehow, in terms of training the trainers, we need our teachers in Wales—I understand that working-age skills issues shouldn't be forgotten—to be absolutely cognisant, on board and incentivised to deal with this and pass that learning on to kids, because otherwise we've had it for the next 60 years, potentially.

Do any of you disagree with anything that any of your colleague panellists have said in terms of top recommendations? Calvin said he agrees with everything that's been said. So you agree; great. Do any Members have any points of clarification on the recommendations that have been put forward? No; lovely—. Lee.

Can I ask a final question in terms of what this committee can do? We are currently putting our toe in the water; we've done a couple of panels. One thing that struck me—we obviously haven't had a chance to reflect on this yet—today is how vast the subject is. Do you have any thoughts on what, usefully, we could do as a committee to try and nudge this debate in the right direction?

I would just look at how you could change devolved policy. I wouldn't worry about what businesses are up to; I wouldn't worry necessarily about what communities are up to, even. But in terms of what the Welsh Government can do now to try to move policy in directions that are more—. Leighton has made many points about the economic action plan and so on. What, actually, can you tell Welsh Government to do that is in its own power to do now? There will be many more of these sorts of things happening—committees and otherwise over the next five to 10 years. So, maybe, in opposition to my point about kind of future-gazing 2050, if you were to critique Welsh Government policy right now, as we have to some degree, what would you tell the Welsh Government to do now, specifically in the areas where it's engaged, like agriculture, economic policy and Brexit? And forget about this wider stuff, maybe, just for now.

I would go and see some things that are already happening. You know, a picture tells a thousand words. Go to Silicon Roundabout, go to some local authorities that are actually leading on this and embracing open data and ask them what happened and why they changed the way they did things, what benefits they've had from it. And finally I'd go and actually see some start-ups. PyDiff meets regularly in Cardiff. It's a bunch of guys—mostly guys—interested in tech. Go and see them, have a chat to them, ask them what support they would want from Government and what kind of businesses they're developing.  

Thank you, Leighton. Can I say thank you to you all? I appreciate that it's a lot to digest and a big subject in such a short amount of time. I apologise to those joining us remotely if I've not been able to pick up if you wanted to come in at any point, but I appreciate your time this morning. We're very grateful.

We'll take a 10-minute break and be back just before 11:05.

10:55

Gohiriwyd y cyfarfod rhwng 10:55 ac 11:08.

The meeting adjourned between 10:55 and 11:08.

11:05
3. Amaethyddiaeth fanwl—Awtomeiddio ac Economi Cymru
3. Precision agriculture—Automation and the Welsh Economy

Welcome back to the Economy, Infrastructure and Skills Committee. This is our second panel with regard to our inquiry this morning with regard to automation and the Welsh economy. I'd be grateful if the panel could just introduce themselves for the record—if I go to my left.

Christopher Hoskins. I'm the commercial manager for the SoilQuest precision agronomy arm of Agrii.

Professor Simon Blackmore, head of agricultural robotics at Harper Adams University, and director of the National Centre for Precision Farming.

Jason Llewellin. I work in arable farming in Pembrokeshire, and have been using precision farming techniques since 2006.

Lovely. Great. I should have said this session is also specifically on precision agriculture as well. If I come to Lee Waters for the first set of questions.

Thank you. There's nothing new about precision agriculture and using technology in farms, clearly. Farmers have been using genetics, in particular, for a very long time. I guess the pace of change with digital and technology is what is different now, and there are clearly examples around the world—in New Zealand and America, particularly—where this is being used at scale. Before we get into the nuts and bolts of precision agriculture and how that can be applied to food production and making farms more productive generally, I just wondered about the way that Wales, in particular, can try and get in on the industry, because it strikes me there's a real danger we're going to be importing robots from China and software from America and exporting our data. What more can we do to make sure that this is captured in Wales? There's excellent practice going on. How can we make sure that robots are built here, the software is designed here, and the data is harvested here? What more can Government do around that specifically? 

11:10

Can I start? 

Yes, after you. 

Okay. So, first off, I'd like to say that the topic I'm going to be talking about is going to be very disruptive. So, we are interested all in precision farming. What is the ultimate form of precision farming? It is what I call robotic agriculture. The questions you're asking I think are very valid and particularly valid for Wales.

If I can cut right the way through to the chase to make it relevant to your question—and that is that the big farms and big fields can make use of the big machines, and we've seen that. That's all based on economies of scale and now that's coming to an end because the machines can't be made any bigger. Wales has been disenfranchised from the use of these big machines, because predominantly you've got small and medium-sized farms and small and medium-sized fields. So, they can't make use of the economies of scale. We're now developing a complete new mechanisation system for crop production based on small smart machines, and it's going to benefit the smaller fields and smaller farms more than the bigger fields and bigger farms. So, I think the environmental structures and fields and environments you've got in Wales have a potential. We can't do anything about the climate with the rain and so on, but, in fact, one of the factors about the small machines is that they are ultra light and where the big machines would sink up to their axles and can't go out into the field, and every farmer will then blame the weather, I don't blame the weather; I blame the wrong type of machines, because it's the machines that we design and it's the machines that we build.

So, this is quite radical and there's a lot of new opportunities and a lot of these developments have been made in the universities and research sectors, but my main role at Harper Adams and the new Agricultural Engineering Precision Innovation Centre is to try and commercialise this. But I do see that there is a role for Government, both in Westminster and here, in terms of demonstrating these technologies. These technologies exist. The farmers can't buy them at the moment. But there should be the ability to show these technologies, to show them off, to then help develop them into the commercialisation and then making them ready for full commercalisation. So, to answer a number of aspects of your question: I think there's a new opportunity here, and I think that Wales and the Welsh Government can play its part in not only promoting the use of these types of machines but also in terms of developing these machines for the Welsh environment. 

I do a lot of travelling around the world in this area, speaking, and I've got lots of friends developing all these different things. The technology that we've got in the UK is as good as anywhere else in the world, if not better. We're one of the leading areas. Denmark is pretty good and is leading, and Japan has got some good technologies, but these are all open technologies. We've got the ability to be able to—. Again, to stimulate the development of robots in the UK is something that I'm trying to progress now myself. So, the opportunity is there. 

So, what would you like the Welsh Government to do that it's not currently doing? 

I would like it to work with various organisations like the Agriculture and Horticulture Development Board, the Biotechnology and Biological Sciences Research Council and the Engineering and Physical Sciences Research Council, and so on, who are all asking the same questions and are coming to the same direction in terms of trials and demonstration—getting the equipment out of the laboratories and getting them out into the farms to see how well they work to be able then to do this final part of the development and then commercialisation. 

Lee, if I just bring in Joyce, then I'll come back to you. 

Sorry—? 

Timescales will be important. So, you say that you've got, as I understood it—perhaps I'm wrong—a sort of prototype possibility and you've got to get it from there to mass production, I suppose. So, what sort of timescale do you think that we would be looking at, if we put some direction into it, to seeing an outcome? 

There's no technological reason why we don't have them now. All these things have been developed and we have prototypes. The timescale is just how long does it take to get these things going, to be able to get them into production and so on. I don't think that it would make a lot of sense in terms of the Welsh Government putting huge amounts of money into this, but it just needs to be the pump priming, it needs to be the catalyst, to make this happen. If we look around the world at investment in this area, the main area is in California, and it is hundreds of millions of dollars that are going into these technologies now, these new ag-techs. And, of course, the UK Government has been very active in terms of ag-tech. There's a lot of money in this sector, and it's then been promoting it and developing it. It is a case of being able to de-risk some of the investments that are needed. I think that's the role for the Governments: to be able to promote these or set up these schemes, allowing these different technologies from all the different universities and research sectors to be tried out. It might even take the form of some form of Grand Challenge, but there is the need there to be able to push this forward just that next half step to be able to attract the significant funding that we're seeing.

What happens in the States, of course, happens here not long after. So, the big investment in ag-tech in the States is starting. More and more investors are talking about this from outside the ag sector, so this is real good news. There's a lot of money outside that people are prepared to invest, but they just need to be able to see these machines working in a more realistic way than we have them at the moment.

So, to answer your question, 'How long?'—we're working flat out on that. I think, with the companies that I'm—. I've got three start-up companies working on these topics now, so they'll be doing field trials in the autumn. So, within a few years, hopefully.

11:15

What are the benefits in terms of environmental or animal welfare?

Well, I'm sure my colleagues here will tell you about the advantages and opportunities within precision farming. It's all about understanding the spatial and temporal variability, but it usually uses technologies to be able to achieve that. So, when we then have micro-droplet application on weeds—. So, there's a lot of discussion about whether glyphosate should be banned. I say some pretty radical things. There's nothing wrong with glyphosate. We should ban the 36m boom sprayers that apply the glyphosate in the wrong areas. Nobody's talking about that. I'm looking at this from a systems point of view. We can now use machine vision to recognise 26 different species of weed. We can then apply the chemical only onto the leaf of the weed. So, we're not putting the smarts into the chemical anymore, needing new active ingredients, new selections, and we don't need new genomics or anything like this. We've got these machines. These are the machines that we have, we control, we can build; it's non-disputable. We can just put chemical directly onto the leaf of the weed, and we can save 99.9 per cent of the chemical straight away. If we go to the next stage—. Sorry, colleagues, I'm dominating the—. I'll quieten down in just a moment.

Another project we've just finished is called laser weeding, so we can use the cameras to identify the weed, so, identify the meristem of the weed, heat the meristem up to 95 degrees Celsius, cell walls then rupture and the plant then either goes dormant or dies. It takes about 7W of energy to do that—no chemicals at all.

Perhaps if I ask your colleagues in terms of benefits for, particularly, animal welfare.

Well, we're arable, so we're not in and around animal welfare, but, from an environmental point of view, we've seen less run-off and less erosion from when we've applied more seeds in poorer areas, because we've had a better crop. We've also felt that that's absorbed and used the nutrients more efficiently, and then also by doing our variable spreading of applications of inputs, we're actually putting the fertilizers where they're needed, where the soil is poorer, and reducing in areas in which it's good. So, we've seen crops become more consistent. And, I've no evidence to say this, but I believe that the water quality is better, because we're putting it where it's needed. So, in an environmental thing, I think there are definitely soil and water benefits from using precision technology.

I guess I come at it from a similar level to Jason, in that we're supplying the software that Jason's using to apply, so we're using various bits of technology to manage and measure crop performance, and then apply accordingly. So, as Jason rightly said, whether it's looking at soil type, looking at erosion factors and putting more seed in those areas, putting the right nutrients where nutrients are needed, and getting the best out of those nutrients—. And we're using various levels of technology to do that. Some of that is relatively expensive technology in terms of the GPS equipment that's required, but some of that technology that we're using is very readily available to everybody, whether it's arable or livestock, using things like satellite imagery, which is high-resolution imagery and relatively economical to all parties, and those are the sorts of technologies that we're using to really help monitor crop performance and target nutrients in that way. 

11:20

We're also using guidance systems as well so we're not overlapping. So, we're potentially using less fuel and autoswitching on machines so they're not overlapping, so we're using less of the pesticide products and putting them where they're needed, not double dosing, which, you know—. Those sorts of benefits, really. 

David Rowlands, do you want to come in with your point? 

Yes. Simon, you've largely answered the questions that I was going to ask because, obviously, Wales from the point of agriculture is pretty topographically challenged, isn't it, but you've actually said that this smart technology can actually help the smaller farms, et cetera. So, I just want to move on a little bit, then, from that, which is interesting, to perhaps the hardware that we're going to be using that will make it easier for hill farms. We know the GPS systems. It'd be interesting, Jason—what sort of farm do you actually farm? Is it a hill farm, or is it—  

No, it's lowland. It's rolling, but it's—. We have some hills, but we're not hill—we're down. So, we're able to make use of the technology very successfully, but I don't believe the sort of—. We use relatively big equipment in relatively small fields, so there are some benefits as well in the smaller fields, because we have less overlapping or we've got more accurate boundaries that protect the boundary. We are using precision applications with the fertiliser, which drop the fertiliser straight down, so it's not going into the boundary, it's not going into the margins, it's not going into the hedges. So, there's a lot of—. The technology can be used on all sizes of fields and farms, I believe. 

I think the perception has always been that precision farming is very expensive because of all the various GPS boxes, but if you look at some of the software that we're launching now, which, as I say, is readily available, to utilise that technology, not necessarily to a precision level as we know it, but as a management tool, all you need is a smart phone and some level of 3G coverage. And that's all you need to supply you with very, very high resolution, local weather data, satellite data, the ability to actually utilise that out in the field, ground truth it, measure it, test it, and then make management decisions based on that data. That can all be done very simply from a smart phone, or anything with a GPS SIM as well as a Wi-Fi-enabled laptop computer.   

I think one of the important things would be, if I could just follow up on that, coverage.

All of these technologies require connection. And there's a discussion about 5G now. So, if there's any opportunity to get better coverage—.

If there's a barrier in the way to it being used, it's reliable, good broadband and 4G, really. We're very lucky; we've got a new mast that's gone up on the farm, so we now have an amazing 4G signal, but, yes. 

It's very interesting to see that most of the technology seems to be about crop production. Do you see anything, Simon, with regard to, obviously, animal production? We're talking about the topography of Wales. It's largely sheep and some hill cattle and that, that we're talking about. Where might it help there?  

I think the technologies are more applicable in certain sectors. Yes, you're right—I tend to look more at the crop production side, and that tends to be in the high-value row crops and the vegetables and so on, because they have the higher value. In terms of animals, there are very few technologies being used with sheep as far as I'm aware, but robotic milking is now well in the mainstream, and, in fact, what we've seen is that it's as much a social issue as an economic or any other issue, because where you're getting the family farms and the young person is then deciding if they're going to stay on the farm or go somewhere else, then the idea of having robotic milking—even just a single day on a small farm—is becoming quite prevalent now. And so the use of this technology is enabling people to remain in the rural areas rather than having to go elsewhere.

11:25

I don't see why what we're doing on a crop basis can't be utilised in grass production, which then ultimately is into the animal sector. I think the utilisation of nutrients and soil mapping could have just as many benefits in a grass system as an arable system, as we're doing. 

I had a colleague talking about slaughtering on farms and then transporting carcasses off the farms rather than transporting livestock off the farm. So, obviously, smart technology could come in with regard to the monitoring of that, couldn't it, and how it's—.

Tracking and tracing is an important issue. As consumers, we all want to know where our food has come from, and has it been treated in the right way, and is it of good quality. From the precision farming point of view, we've got the ability now to be able to locate, particularly for the row crops, each individual plant with this thing called the RTK GPS, which measures the position down to a couple of centimetres. And so we can know exactly where that's come from and what has then been applied to it and what hasn't been applied, and we can track all the way through then. One of the growers that we're talking about at the moment, they say, 'Well, can you go with your smartphone to the supermarket and then say, where has this come from, what has happened to this?'

How reliant are you on superfast broadband coverage in terms of the technology that's emerging?

Well, all of the—. We've got a trial on at Harper on at the moment called the Hands Free Hectare and we've just started a project where we're involved in 5G coverage to be able to see then what band width would actually be needed. And it's really just to trial it out. So, I think there's four or five HD cameras on the tractor, and we're just trying to push it to its limits. But the connectivity, 4G connectivity—3G, you can't really do very much with, but 4G, when you've got the 4G, then—. And all the machines I'm ever talking about have got to be connected to at least 4G all the time. 

So, what are the consequences in terms of what we're talking about if we don't have good 4G plus?

You can't use the technology. 

You're disenfranchised straight away. 

I just want to come in very briefly on the reference to meat production. One of the frequently cited barriers to export of particularly lamb carcasses to the middle east or more widely is shelf life. I understand from correspondence I received from Welsh Government a few months ago that currently the only place in Wales that can match New Zealand shelf life is the Rhug Estate near Corwen. What role can precision agriculture play therefore in helping lengthen the Welsh shelf life of meat carcasses?

I have no idea. 

Anybody else? No. That's fine. We'll put that to somebody else, that question. 

Yes, I'm quite interested in the project, the Hands Free—

The Hands Free Hectare. 

Yes. I think it's using existing components largely, so that you're free from the large vendors or whatever. As the technology and the market matures, do you see the normal trend with technology that some of the larger companies will get heavily involved? And are there network effects with the data, so, basically, there's value in the data that's being collected? And who owns the data? Is it the farmer, or is it the technology company? 

I think there are three questions there.

The Hands Free Hectare is a technology demonstrator using classic equipment that we all recognise, and then just putting free open source software on it just to show it can be done. As I say, this is robotics, not rocket science—it's straightforward engineering.

I've missed the second point. The third point was data, and we were just having this discussion—.

11:30

The second point, really, is related. At the moment, it's smaller machines, but also smaller start-ups, but do you see the larger companies, basically, weighing in?

I've been badgering the John Deeres, Massey Fergusons, CLAASs, CNHs for 15 years in this area—not interested, not interested, will not invest. 'Why should we invest money in disrupting our own business? We've got a nice, linear development here—the next machine's going to be bigger than the last one—so, we don't need to do it.' So, this is why, to the earlier question, I was excited about money coming in from outside the agricultural sector, and it's going to be the start-ups that are going to have the major impact now, because they've got no legacy. We can just look at the whole agricultural process, come up with the machines that are best suited to do what it is the farmer wants, and use the technologies from outside of the agricultural sector to bring it across and in. That's why I'm saying it's going to be disruptive. But the big companies do have a buy-out of that, in that they have growth through acquisition. So, when I was in Denmark, I had some students set up a company, and CLAAS bought the company and incorporated it into the tractor manufacturer.

And then the third question: who owns the data? Will it be the technology suppliers, the companies, or the farmers, or both?

As you say, we had the conversation earlier on, and my understanding is that it should be the person paying for the data who owns the data. So, from our perspective, we cover a large part of the UK with satellite imagery, and the satellite imagery is owned, if you like, by the individual grower who's bought his pocket of data. As a business, we own the entire data, but we're not selling that data; the data is there to be used with our own research and development models, to improve the research and development models, and the growth modelling, and all the other things that we use that data for internally, rather than selling it. But it is a significant risk at a farmer level as to what else is going to be done with his data.

A related question, in terms of the ability of these larger companies—the John Deeres, Monsantos—to manipulate market prices, based on the data they're receiving. So, if they know what particular crop is going to work well in a particular climate, they then have that data to know where to set the market price of their selling. So, what thinking is going on, what's the feeling about how that manipulation can take hold?

This comes back to this fundamental question of who owns the data, who has the right to manipulate it and who can have the ability to cash in on the value of that. There are some big companies that have been set up recently that are just purely into this big data side of things. One of the four new centres at Rothamsted—what's it called, the big data centre there? Agrimetrics. It has been set up there to be able to deal with this. Who has control of it? There are issues in terms of—. I agree that who owns the data are those who pay for it, but as we, again, were talking earlier, when the satellite is taking an image of somebody's farm, and they have no say in how that information is being used, and somebody is there making money out of what is going on on somebody's farm, there is an ethical issue there. There must be some right to be able to control that, or have—

To relate this back to the session we had earlier, you can see that the line is that data is to the fourth industrial revolution what oil was, or coal was, to the previous ones. And I've been reading about Nike, for example, in that modern trainers generate clouds of data—so, how you behave, and so on—and that is worth more to Nike than actually selling the trainer. So, it may be that they'll give away the trainers for free to own the data that is generated from them. So, applying that to farming—.

Exactly. The exact thing happens in farming, where we've been looking at some of these row crops—broccoli; the growers have to grow 50 per cent more of the broccoli, and they're still throwing away 20, 30, 40 per cent at the point of harvest. So, the value is in the data; there is more value in the data about that piece of broccoli than the piece of broccoli that is actually sold from the farm.

So, what are the implications for that, and how do you mitigate it?

Well, we're still exploring that. You're right—we know that there is huge value there, but safeguarding it and controlling it, and who has access to it, is another issue and I don't have any simple answers to that.

11:35

But is there an opportunity, Jason, in creating a digital equivalent, if you like, of the old agricultural co-operative? Farmers who are early adopters could come together and say, 'Why don't we collaborate and share the data and share some of the value in that?'

I think, as farmers, you need good back-up to get this stuff to work, so we need the back-up of the likes of agri and the GPS companies. So, as a farmer, I feel you're sort of trying to get a hold on a lot of information to cover a lot of different subjects and not specifically into the GPS and the data—you need to make use of it and you need to be able to interpret it, but I feel that there's not a great uptake in Wales of even what we're doing, and I think there's a long way to go to make that a lot more useable and used. There are things that, in the way the Government helped and subsidised different things, could be changed and could be done better to encourage it. If you've got other environmental issues, like nitrate vulnerable zones, or if those sorts of environmental things come along, this is a good way of reducing the potential pollution, but also you've got proof of what you're doing and you're accountable for what you've done so that you get the—. Precision farming is a good tool for that as well.

So, adopting these technologies is more critical at this point than the question of ownership, which may be for another day.

Yes; possibly, yes.

I want to move on to a new subject area of jobs, skills and employment and keep in mind, in your answers to questions, that we want to know what Government needs to do differently. I'll come to Vikki and Joyce wants to come in as well. Vikki Howells.

Thank you, Chair. I'm curious as to the degree of upskilling that's needed, if we're going to really embrace precision agriculture. Most of the things I've read suggest that the sector creates demand for skilled work, but then I've also read some contradictory things about certain parts that require little training, like robotic milking systems and auto steer. So, how much upskilling do we really need?

I think it's not necessarily upskilling. There needs to be training, but there also needs to be time to allow people to accept the system and to use it. If you've got a person who doesn't want to use it, it's never going to work and you're wasting your money. So, I think there is a certain amount of training and skilling that is needed because of the level of understanding of the system. You need to understand the system to make it work—it's not just going in the tractor and off you go. If it goes wrong and you have a problem with it, you've got to be able to try and work your way around that, which needs education really.

I'm thinking also about the barriers to adopting these kinds of approaches, and one of the things that was in my mind is the age of the farming population in Wales. I think the average farmer is about 62 or something like that. So, is that a potential barrier, and what do we do about the new generation coming up? We know that it's difficult to retain people in the farming sector anyway, so what kind of barriers are there? Or am I thinking the wrong way and actually this is an opportunity to get the next generation more fully engaged?

I'll just come back on that one. I don't think that age is necessarily a problem. My dad's 67 and he's quite happy to use auto steer, variable rate—everything. So, it's not necessarily age; it's the calibre, the background and what people are in interested in.

And the complexity of the technology. When we first developed precision agriculture—or the concept—which was more than 30 years ago, the scientists and engineers were quite happy to develop all sorts of different maps and clever things, but the poor old farmer suddenly then says, 'Well, how am I going to make money out of this?', because we'd made it more complex. 

Certainly, we're now looking at what I call, 'leap-frog technologies', and many of us have got phones and these sorts of things. You don't need a PhD to run one of these things. And this is where technology has been developed to such a level of maturity that it meets our needs and our requirements. So, when we're now developing these new types of machines, it then should adopt that sort of level.

But the generational thing is quite interesting, because the biggest advocates that I've had on these small, smart machines are people a generation older than me, where they remember the smaller machines, the smaller tractors, and the ability to go out and work the soil without damaging it. And so, although it is probably the younger generation that is more ready to adopt it, the older generation certainly gets the principles of what is trying to be done.

11:40

And I think I'd back up, from what Jason said, from my experience, when I've been going out with our services, the older generation are more than happy to engage and, in some ways, it makes them feel closer to what we're trying to achieve on-farm, because they get a better understanding of whether it's the variance in the crop that were looking at by engaging in this sort of technology. And certainly, the younger guys take it for granted now, but the older generation, they're certainly not averse to it, from my experience.

And I think we also need to harness that older generation's knowledge of that farm and digitise it. You know, it's in there, their heads, and we need to get it down for the next generation to be able to make use of that knowledge.

So, what does Government need to do in terms of training? Because, from what you've just said, it's almost that they don't need to do anything, really.

Well, the universities are responsive to this, and the colleges understand these issues. The curriculum is updated every few years for accreditation and so on. I'm not sure that it's necessarily any intervention that's needed from Government for this, so long as the training organisations are up to date.

Yes, I think the training needs to be fit for purpose and be more practical, more actually using stuff, so that when they get on-farm—.

Okay. I've got Joyce and Lee waiting. Is yours on this, what we've just said? 

It's around Government. It just strikes me from the conversations I've had with officials and Ministers that the prevailing view is that farmers are just getting on with this; this is what modern farming looks like, so we don't really need to get much involved, because there are plenty of sources of funding, such as grants, Farming Connect, and so on. But I just wonder whether that's strategic enough or whether or not there should be more of a guiding hand from Government, to get people trained, to apply the technology, to encourage the take-up, rather than just allowing this funding to be loosely used.

All I can really think about is to come back to my earlier point about demonstration. There are a lot of new technologies. How do you raise people's awareness? How do you convince a farmer to invest money, their own money, in the new technologies to make them more efficient? And you can go through the normal sales process, but to actually see the equipment and how it's done, being demonstrated, maybe at the educational organisations, would be one way of doing it, encouraging them to show the advantages of this new technology and give the appropriate training so that people can use it, rather than just reading it from the adverts.

On your point about funding, and Farming Connect in particular, it's fairly prescriptive; it's not flexible to encompass different types of farms. Yes, they had it a few years ago. When we wanted to get more precision, we could have some soil tests. We could have 10 soil tests. Well, you know, we're doing two per hectare, so that was quite a lot on 300 hectares and we could only get funding for 10. So, we just got on and did it ourselves; we didn't have any funding towards it.

Same with the GPS kit. Within the small grants scheme now, yes, you can get some basic funding for some basic GPS, but if you've already done that, there's not really a route to go on to the next level. So, it's there, but it's not providing enough options or flexibility to cover the majority of farms and situations.

So, do you think the funding needs to be reshaped to specifically allow precision agriculture to flourish?

I don't think it's necessarily—. Yes, it needs to be reshaped, it needs to be thought about, and it needs to be more flexible.

Talking about flexibility, we are obviously interested in the impact of precision farming or agriculture on employment, because those will be the key questions that we will be asked.

Well, again, there is an interesting answer to that, I think. Where we've seen the increase in the size of equipment, whereas you used to have 10, 20 people working on a farm, these days, then, you've got one or two people. The big tractors have already displaced much of the rural population. When I think ahead into the future about how these new machines will work, the permanent rural population will be the same as it is now, but it will have to be skilled up, as we're talking about, to be able to deal with these smart machines. But the big area that is of interest to many farmers—I don't know how much row crops are done in Pembrokeshire and so on—with Brexit coming, the lack of seasonable labour is a major impact. So, therefore, these technologies that we're talking about—selective harvesting work and the robotic harvesting and so on—will then go to some degree to be able to replace the loss of some of that labour by utilising some of these new technologies. But, as I said, even in my furthest vision, the rural population will not decrease any more but will certainly have a higher value and require more skills.

11:45

So, it'll be raging against the machine, then, rather than people, it seems to me. Moving on again, you talk about upskilling a workforce to manage, sometimes at different scales and at different levels—. Will they be transferable skills if people wanted to come out of farming for whatever reason, because we've got mostly small farms? So, that's one question, and the biggest question for me is the impact on the environment and wildlife therein—if you could talk about that.

Because we're dealing with these sort of high technologies, there is certainly a lot more interest coming into the agricultural sector rather than going out. So, I don't see that there is a real issue there. Sorry, your second point was—?

Environmental. When—[Interruption.]

Shall I bring Jason in for a moment? I'll come back to you, Simon. Jason.

From an environmental perspective, I think, again, from my experience, the environmental impact has been positive rather than negative, because of the targeted placement, the better choice of products. We're not just using basic blended fertilizers, for example. We're tailoring those requirements to best match the requirement of the crop, based on the various levels of data that we've gathered about those crops. So, I think, from my perspective, the environmental side of it has only been increased.

I'd agree totally. Where we're placing the stuff is in the crop and it's not being thrown where we don't want it; we're putting in what we need. We can take into account whether we've applied manures or not and then adjust rates of fertilizers accordingly. So, I think it's just very positive for the environment and I think the soil as well. We've got cultivations that allow you to cultivate maybe a very steep slope across, because you've got cultivations but you can then still spray it up and down because you've got auto-steer technology. That reduces run-off. So, there are a lot of technologies that can improve the environment. 

Once I hear 'spraying', I think of pollinators straight away, which has brought me to my next thought. In terms of precision agriculture, and you're in arable, are you rotating crops?

And have you assessed any impact, albeit negative or positive, on pollinators?

We have a six-year rotation that we grow. We have cereals every other year and we have a break crop in between. We have oilseed rape every six years. We have beans every six years. We have two local bee farmers that take hives to the fields. We've got them there now. We resist the use of of insecticides immensely and, on certain crops, we don't use them. We believe that our biodiversity is increasing. We have quite a lot of margins about the place, which, again, we've been able to manage properly because we have a defined boundary within the GPS systems that don't encroach on those, which help our biodiversity. So, I think so, yes.

11:50

I'll come to you in a moment, David. I just want to move on to unintended consequences and barriers, both of which the Government need to address in regard to precision agriculture. David, do you want to ask your question?

Yes. I just want to go back a little bit to Lee's questions with regard to Government intervention to increase take-up. Chris, you're at the cutting edge of introducing this agri-tech to farmers. Is it a hard sell?

The way that we've currently launched it, because we—. Anything you launch that's new is always going to get taken with a little bit of caution, so we've actually launched it as a free trial to really get people to engage with it. And we've also launched it in such a way that when it goes into a paid service, it's at the very, very low end of the cost spectrum. And we've done that deliberately to try and get as many people onto the system, to get as much uptake on the system. That, in some ways, gives us more data. It comes back to what we were saying about data, and, from our perspective, the more data that we have—. Part of this system that we're promoting now is around growth modelling, so it allows us to predict the yields of crops and the biomass of crops and the growth stage of crops using various different satellite technologies. So, by offering it to a wider audience at a very, very low-level cost, it allows us to gather as much data as we can to improve that model quicker going forward. 

Can I ask, in terms of barriers, what are the barriers to precision agriculture? Give me the headlines, not the detail. 

We covered it a bit earlier. Broadband and 4G is quite a barrier.

Broadband is one. Broadband and mobile. Okay. Is that the top barrier?

I'd say it is, yes.

Can I just jump in there? I accept that there are parts of Wales that have 4G and the farms in those areas aren't taking up precision agriculture to the extent that they could be. So, of course it's a barrier, but that's not the only barrier, because even when that is present, you still aren't having these technologies adopted.

So, I'm asking what the top barriers are. We've addressed one. What are the other barriers? In headlines—give me the headlines, not the detail.

There is a bit of an issue with precision farming, in that people tend to think of it as the technology—you buy the technology, you're a precision farmer and you're going to make money. And that's not the case. Precision farming is between the ears. The tools of precision farming are allowing you to do something new in a different way. Whether you make any money out of it or not is entirely up to you. So, it's never a clear case, a clear benefit by adopting precision farming. But the process—again, we were talking about this earlier—and when people go through the process of looking at the farms in more detail, trying to understand what's going on there and do something about it, that's an invaluable asset as it makes people look at the farm in a very different way than what they did before. And, usually, it is for an environmental benefit and sometimes, and probably often, it's an economic benefit. 

It takes, I'd say, a lot more, but there is a certain amount of extra management time needed. So, when everybody is busy or working against the weather, it is something extra to do, but I think there are good benefits to be had from making that investment in time.

So, tell me, succinctly, what are the key barriers. So, I think you're saying finding the time, in terms of prioritising. But broadband, mobile—what are the other key barriers?

I think perceived cost has to be up there as well, because a lot of people perceive it to be very, very expensive. And it can be at the extreme of real-time kinematics or at the management level that we're promoting at the moment—free or at an 80p an acre type cost. That's a very, very different conversation to be had. So, there is a perception that it is expensive.

But it doesn't have to be.

It doesn't have to be. You can spend as much as you like, but you can also just do simple things relatively cheaply.

11:55

Do the banks understand the opportunities that exist in terms of precision agriculture and lending money to farming industries?

Again, I guess one of the other barriers that we will always come across is how you try and quantify the advantages of precision from a financial perspective. We've talked about many of them today, but to actually quantify it to say, 'That element of precision farming has saved me, or increased my margin by, x' is a very difficult figure to put.

Why would that be difficult, because obviously if you're using far less fertiliser, and possibly going to be using infrared, that must be a cost-effective—?

One of the things that you can do with precision farming—very rarely can you ever increase the yield, but you can significantly reduce the cost of production. But the point is: why do farmers overapply fertiliser and chemicals now? It's to reduce the risk. So, by cutting back, you're potentially increasing the risk, but you're mitigating that risk by having more information. So, it's a balance between the two. That's why it's—.

You don't always, in our situation, apply less. You may be applying less in this area, but more in that area. So, the average, within the mapping that we've done—. So, within a field, you're doing it in different places. So, not necessarily, but overall you're evening your crop up and putting it where it's needed, which is of environmental benefit. 

Thank you. Professor Blackmore talked about the impact on the population. You said that the rural population should not shrink, provided that the population can skill up to meet the needs of the new technology. Mr Llewellin, you spoke about biodiversity benefits, but what are the threats and opportunities in terms of farmers spending locally, accessing services locally, but also in issues such as food tourism, environmental tourism, and 'keep food local' campaigns?

Well, I think the opportunities are there in terms of—. A lot of these technologies have been developed outside of the agricultural sector, so we're bringing them across—like GPS, RTK GPS and so on. So, the opportunities are there. The threats—certainly, from the fundamentals of precision farming, it's better for economics and it's better for environment and the two are actually pulling in the same direction. Normally, if you come up with a new technology, you can make money but it's bad for the environment. But the two are actually pulling in the same direction, and the concept of the small, smart machines is just taking it to the ultimate: how far can we go with technology? 

If we look at the type of threats, maybe investment, maybe some people will think that maybe the robotic systems will be expensive. I don't think that they have to be. They replace people in the workforce—only those, I think, in seasonal labour. There are various—. You were asking—

We discussed that earlier—putting farmers in the hands of multinationals to manipulate prices. That's a threat.

One aspect that I wanted to bring up, and hasn't come up, is legislation. So, at the moment, perhaps you're aware that everybody's flying drones around these days. We can use drones on farms—spray drones. We've been doing tests on different types of spray drones at Harper. You're not allowed to use them. So, if a farmer—my colleague here—wants to say, 'Right now, I don't want to go and spray 36m; I want to go and spray 2m of weeds just out there with a spray drone', he can't do it because legislation at the moment is saying that you're not allowed to do that.

But there are reasons why that legislation is there, so what do you think should happen?

A lot of the chemical regulations directorate is all based on hazard, and it needs to be changed to risk. So, if he's only going to go and apply a couple of grams of chemical to that one patch over there, the risk of anything going wrong is very small, even though the chemical might be seen as being hazardous. But, in China, talking to a company recently, 100,000 farmers are now using spray drones on their fields, and no farmer in the UK can use spray drones. So, there is a range of legislative issues that need to be addressed to allow us to then take advantage of these opportunities.

Anecdotally, in New Zealand, the reason that they can bring sheep up is that they fertilise the hills—it's done by crop spraying. Obviously quite an expensive operation with an aeroplane and a pilot, et cetera, to do that, and that's what I was thinking: that this sort of technology, using drones, particularly with the topography of Wales, again, would be—.

12:00

Well, certainly the smaller machine, yes. I think the problem in New Zealand now is that where they are doing this mass coverage, it's causing more problems than they're prepared to accept, but when we are then targeting just down to much smaller areas, then I think that's a lot more acceptable.

We're drawing to the end of our session now. Our piece of work will result in us perhaps making recommendations to Government, so what do you think our recommendations should be?

If we want to take advantage of these opportunities from new technologies, legislation framework must be in place to not discourage, but encourage, the use of it. As we move into new technologies, there is always a risk associated with that, and I think that Government can play its part in de-risking this or demonstrating or knowing the risks where a farmer, perhaps, is not prepared to invest yet, until they see this, sort of, being demonstrated.

Give me a bit more of a specific example of what you mean.

Well, what I've been recommending is that we have—. As I say, you've heard of the hands-free hectare; we're now developing the concept of the hands-free farm. That's not necessarily just using the existing equipment, but it has the capability of using new equipment. A scheme could be set up to fund or partially fund or co-fund one farm in north Wales, one farm in south Wales, or different sectors, where it could be offering the framework to the universities and research sectors to come and trial this work or to demonstrate these types of machines, or even private industry to actually come and show how it can be done. But it's to take the risk out of it to allow private money to lever up that initial investment.

Okay. And the same question to Jason and Chris, but I'm thinking of what Government need to do from a strategic perspective.

I think Jason touched on it earlier in terms of the variety of grants that might be available. It seems—and my knowledge of it is less than Jason's—but it does seem to be that it's quite specific in certain areas. We've talked a lot today about things that are happening above the ground—and I'm conscious that we're almost out of time—but there's also a lot of things that are happening below the ground that we haven't touched on today.

Well, things like soil health, soil biology, organic matter levels, all of which are critical to optimum crop performance and have environmental implications as well, but it's not something that we do a huge amount of or isn't probably widely done an awful lot of.

Well, I'm just thinking, in terms of granting and funding, yes, we talk about grants towards GPS boxes and fertiliser spreaders with GPS capability, but maybe we need to start looking at grants towards more basic elements, whether it's basic soil sampling or soil health auditing—things that are happening below the ground that are all integral to precision farming in general, but just general soil health, and I don't know whether the flexibility is quite enough at this stage to cover that sort of thing.

I think, yes, those sort of points, and we touched on training: there's a certain amount of training and just general education in the next generation that needs to be done, but also any grants or encouragement that farmers need, it needs to be flexible. No two farms are the same, and having a prescriptive grant scheme, it doesn't fit very—. It fits in certain circumstances, but it doesn't fit to cover the majority of farms, I don't feel, and—

But if Government made grants more flexible, would that result in a move to precision farming?

I think there could be more uptake in precision farming, and, from our experience on the farm, I think—like we've touched on—it'll have environmental benefits, and it could go some way to helping some of the issues that are out there currently today.

My worry in just making greater flexibility without any strategic focus to it is that the odd farmer like you, the outliers, the innovative farmers will do it, but, overall, the sector won't really take it up because there are other barriers, and without a strategic approach to address all of the barriers and create a framework around that, we're never really going to get the step change we need. What do you think about that anxiety I have? 

12:05

I sort of accept the—. I think part of it as well is that we've lost the people on the ground, to a certain degree. We used to have quite a good guy within what would have been the Environment Agency, if you go back that far. So, we've lost those people that were maybe just helping farmers to encourage them to do small things as well. So, if it was a benefit perhaps of precision technology to help the environment and, 'Yes, you can apply for the grant, but you just get this bit of box, but you don't necessarily get any help with it'—. Going back to my soil sampling example, that was a classic one of it not fitting; it fitted one farm but not the majority. 

Lee's point is that not every farmer is a Jason and having wider flexibility on grants isn't going to—I don't know, but from what's been said this morning—address the issue of moving towards precision farming. 

I'm not sure that any of us have got the answer. 

But there is one other topic that might be of interest, and that is before I came back to the UK, I was running the European FutureFarm project. And that was about an information system that would then relate from government to the farmer directly and be able to go backwards and forwards. You probably haven't heard of it, because I couldn't find a UK partner at that time, but if you look at the reports on www.futurefarm.eu you will find that there is a set of tools there that might be of interest to the Welsh Government to be able to relate to farmers in terms of information and big data.

Okay. In that case, can I thank you for your time this morning? You will be sent a copy of the proceedings and, by all means, review that, and if there's anything else you feel that you want to add to this morning's discussion, then please let us know. But we're very grateful for your time this morning. Thank you very much.  

Thank you.  

Thank you. 

Thank you. 

4. Papurau i'w nodi
4. Papers to note

So, item 4, there are a number of papers to note if Members are happy to note those. 

5. Cynnig o dan Reol Sefydlog 17.42 i benderfynu gwahardd y cyhoedd o weddill y cyfarfod
5. Motion under Standing Order 17.42 to resolve to exclude the public from the remainder of the meeting

Cynnig:

bod y pwyllgor yn penderfynu gwahardd y cyhoedd o weddill y cyfarfod yn unol â Rheol Sefydlog 17.42(vi).

Motion:

that the committee resolves to exclude the public from the remainder of the meeting in accordance with Standing Order 17.42(vi).

Cynigiwyd y cynnig.

Motion moved.

Under item 5, under Standing Order 17.42, can I resolve that we exclude the public from the remainder of the meeting? Are Members content with that? Yes, thank you. 

Derbyniwyd y cynnig.

Daeth rhan gyhoeddus y cyfarfod i ben am 12:07

Motion agreed.

The public part of the meeting ended at 12:07