Welcome to this second panel in today’s wrap-up hearing on artificial intelligence and the impact on the UK and our workforce. We are really grateful to you for coming along this afternoon. The theme that is going to run through this hearing is whether our workforce, our young people and our older workers have the skills and capabilities needed to ensure that we, as a country, thrive in this new economy. We are especially interested, as workers get into later life, in how they are going to reskill and potentially acquire savings to help them do that. Are the Government’s entitlements to retraining in the right place? Are people’s expectations in the right place? We are going to start at the beginning of people’s learning journey.

Thank you for this opportunity. We are pointing in the right direction, but pace is too slow. The CAR—the curriculum and assessment review—took place. Those reforms are welcome. The changes that DFE has accepted in terms of computing recommendations go further than that, broadening the computing GCSE, and adding AI and underpinning data science, including bias in AI, to the curriculum. There are clearer digital literacy expectations for every student up until the end of key stage 4. The timetable is too slow and out of sync with the pace of technological advances. If you imagine a year to spec out a new curriculum, a year to write the corresponding qualifications, a year to onboard schools and select those qualifications, two years of course study to get your GCSE and another two years for A-level, suddenly we are at seven years from now for those young people leaving the formal education system. My big message is that we need to find a way to inject a greater level of pace.

To check this, you are saying that it is going to be 2033 before we have the curriculum in the right place.

No, the curriculum will be introduced in 2028 with the DFE plans. If you think about those young people going through that, taking the two years of GCSE study and then potentially an A-level, T-level, V-level or whatever, after that there will be the 2033 cohort of 18-year-olds who benefit, yes. That is a big watch-out. We need to think about how we make these changes in a way that does not rock the great education system that we have, the great entitlement for all and the good-quality national qualifications that people generally understand. One idea is a core common spine, but with enough flexibility for rapid and culturally recognised iteration.

You probably need to be a bit more ambitious about the pace.

Give us a sense of that.

In the response to the CAR recommendations, there is a response that says that you will look into pace and maybe updating more regularly. The last computing curriculum was written in 2012 and it is only just being looked at now. We cannot be there again, so we need to look at it annually at least.

Let me make sure I absolutely understand it. The 18-year-olds coming out of school today will have rolled off a computing curriculum that was last updated in 2012.

Yes. That is correct.

Yes, mine mainly. It is not impossible. It is difficult with the current system, so we need to think about system change. We need to think about creating space for contextualisation. We need cultural change within the education system to recognise that we are creating the forward-facing future population to thrive in this AI and digital-enabled world. To do that, we need to not teach about DVDs and Blu-ray any more, but to teach about AI and vibe coding.

Are we still teaching young people about DVDs and Blu-ray today?

In some subjects, yes, such as media studies.

There is a lot of what Julia said that I would agree with. I am a computer scientist. I was a computing teacher before I moved into trust leadership. For sure, the current national curriculum is out of date. There is a real risk that the next one does the same if we are not careful. I would caution against the urge to over-prescribe what the new national curriculum is. One of our biggest concerns is that the programmes of study become too over-prescribed. We have schools doing a really brilliant job of teaching digital literacy and AI, which are tapped into what is going on in their local communities. That is based on the experience of children and what families are telling them about their children’s experience of technology. That is guided and supported by the trust and networks that people are part of, such as your organisation, Julia, and charities that operate to provide pupils opportunities to code. From my perspective, we do not offer a great number of qualifications around computing and computer science. Our focus is on building powerful knowledge across the whole curriculum, developing pupils’ broader skills, critical literacy, their ability to think increasingly through systems thinking and their ability to speak well. We recognise that those skills are the primary skills that employers are going to want ahead of, perhaps, the ability to code when you can vibe code. I do not want to undermine the sense that we want to understand the underpinning technology. Young people need to know how this technology works, but they do not all need to be able to get a GCSE or an A-level or pursue a computer science or AI degree. I am more relaxed about the time it is going to take for qualification reform. On the national curriculum, I would encourage us to be as lean as we think is sensible, to enable schools and trusts to use their local labour market information and broader context of the children and communities and what is happening now.

I agree with what both Lauren and Julia have said. The key to this is that, as was articulated by Julia, you have a system that clearly gets the pressure of quality, consistency, recognition and so on. I am not an expert in the school system, but you see the same in the apprenticeship system, and even in the degree system, to a degree. Then you have this alternative dynamic, which is very strongly visible in AI, of rapid development of technological evolution, which is on a different clock speed, to use computing terminology, from the education system. Lauren put it very well. I have an example. I was up last week seeing Amey, the big construction company. I asked, “Are you guys finding AI coming through in apprentices, T-levels and everything?” They said, “It varies depending on who we are getting to do the training.” Lauren’s point is that, in many cases, trusts and others are doing this because they are good trusts and they think about the bigger picture and so on. The balance we have to get to is how you bring everyone up to that level and exchange best practice, get people lean and dynamic, and get people, particularly the educators and maybe some of the curriculum structure, better so that they have the space to do this. It is a difficult balance that is a struggle. In previous iterations, the cycle was probably acceptable, generally. As digital technologies come along more strongly, it has become more difficult. It is an issue we see everywhere, but it is definitely one that we need to find our way through. We will talk more about it, I am sure, as we talk about adult skills.

It has been brought into very sharp relief by the Milburn review, but had previously been discussed fairly widely in terms of the challenge of young people being able to access the system. I would not necessarily lay all that blame at AI’s feet, which of course is a general trend at the moment, where people are AI-washing, or whatever the term is. They are using that as a way of articulating maybe a more fundamental challenge in the labour market. We have seen a lot of changes. There are the million young people and the skew of the apprenticeship system towards older rather than younger people, but also the social dynamics of people coming, post covid, to various other areas and finding them difficult to access. The evidence around whether AI is consciously stopping young people accessing the capabilities is difficult to determine fully. There are some studies that say it strongly, and others that contradict it. We owe our young people the opportunity to get the skills that allow them to be ready for work. We see lots of dynamism in that.

I think the answer is that we do not know. Is that the bottom line?

I was going to say that the biggest thing we get from people is, “We want skills other than technology, because technology is levelling things.” To your point, it is very difficult to say definitively, “Yes, it is because of AI that suddenly we are not getting young people.” I do not think that that is, by any means, the whole story.

There is definitely the need for young people to have what were traditionally termed softer skills—so, more employability skills, alongside the technical skills. There is sometimes perhaps a misalignment in expectation from employers about what an entry-level job should include. Perhaps in some ways it is more senior than it traditionally was, with AI in some places taking on some of those tasks associated with a fresh graduate role, for example. One interesting point to make is that we know that undergraduates really benefit from employability skills. When they have a placement or internship as part of their course, they are more employable. It is not as well known that, when employers take on placement students, they are then more likely to take on young people, even young people they have not met. Their attitudes, culture and hiring practices shift. There is something there, in that it needs to work both ways. We need to support, encourage and incentivise for it to work both ways, so that employers are actively seeking them. These young people have a lot to offer. I negate the term “digital native” because—you know, no. They are certainly attitudinally ready and up for developing their role with AI and being able to add value to the business. Thinking about changing attitudes of hiring employers as well is important in all this.

Turning to adults, do we think the current adult skills and training system is fit for the future, thinking about the evidence that we have heard about AI and digital skills, but also about the leadership and management competencies, which the evidence this Committee has heard has indicated are so important for understanding what might be possible in any given business setting?

The system is evolving, there is no doubt about it. If you were asking whether it is completely in place now, the answer, to take the Chair’s view, is that, no, it is not appropriate to do that today. It is evolving because we are providing short course capability, for example, and so-called apprenticeship units. If you look at the work we have also done with the technology industry and the Department for Science, Innovation and Technology, in that environment we are working with vendors to take their capabilities and provide them to young people, and to people generally, frankly, with the capabilities around it, to say, “This framework assures a certain amount and we can give it a certain badging to give it some credibility.” There is a lot of work going on. There is an AI unit. There are AI apprenticeships that have been created on fast track. There is AI in a number of the standards already, but there are 700-odd standards, so there are a huge number of them. They have a level of stability, as was described before in the curriculum context. We need to evolve, but of course you do not want to throw the baby out with the bathwater and crash apprenticeships, like what happened before with the levy. It has a long way to go, to be quite honest. Part of it is the discussion you were having before about the whole system being more joined up. There is quite a lot of fragmentation, even in the AI space, today inside Government. That needs to be pulled back together if we are talking AI specifically, and then bringing in digital skills more broadly.

On the leadership and management point, do you think that the work Skills England is leading is doing enough to ensure that leaders and managers in enterprises understand what the potential for growth and development is?

If we are specifically talking about the things we are doing within the formal structures, there is a lot of work going on to do that. There are funding changes that have been made around leadership and management, on the basis that the apprenticeship system had skewed itself strongly towards older workers post 25. Difficult decisions had to be made in those circumstances as to whether you move people past. We need to disconnect that from any narrative or intent from ourselves, let alone Government more broadly, that leadership and management are not really important. Prior to any of the downturn, or even the uptake of management and leadership in apprenticeships, people were teaching leadership and management in those circumstances because it is an important thing for any business to do. When I ran a business, I had lots of leadership and management training capabilities because it was the right thing to do to develop my workforce. This sense that, “If Government don’t do this, we don’t do it” is flawed. A bit of that has crept into the system now, which says that, if it is not being funded as an apprenticeship, it is not a good thing to do. It is for the business to make that decision, and it would be wrong to make that assumption.

If employers were here, we know what they would say about apprenticeships, and we can put some of that to the side for the moment. We know their criticisms of the levy, the system and so forth, and their difficulties with apprentice wages. Would they say the content of the apprenticeships on offer is keeping pace with their business needs?

It varies. I agree that there is a lot of noise from employers in there. Julia mentioned it earlier: the employers that take on either young people or apprentices are incredibly convinced of the value of apprentices. Amey, again, is an example. I went to see XL Motors up in Coventry the week before. There are lots of them, and they completely buy it, but I also agree that there is a bit of a hurdle to get over. However, if they then said, “Is that apprenticeship in”—I don’t know—“plumbing now preparing me for AI?”, the short answer is probably no. There are short-course AI upskilling modules available, which are 30-hour modules, which can be funded by the levy. There are three of them that have been created. That is a start on that. Also, if you look at the AI boost programme, which is the commitment from Government to train 10 million people, that is an available, free-to-use curriculum or capability that can be used to supplement things. I was in No. 10 this morning with a bunch of people talking about AI. We have had three meetings in that. It was quite obvious that the things that people wanted to teach three meetings ago are different from the things they want to teach now. Three meetings ago, they were talking about prompt engineering and so on. They have moved on from that now. It is about agent management and context management. Trying to keep up with the technological advances here in some sort of Government context, frankly, is impossible, but Government need to put the framework in place to allow people to do this in a more dynamic way.

This just-in-time training is probably going to become the model. Trying to keep up is, as Phil says, probably fruitless. At BCS, we run an annual tech priority survey among our members. We ran it in 2025, and the Committee already has that research available to it. We have a live 2026 one running at the moment, so I have hot-off-the-press information. It is unsurprising to know that AI skills is the top in-demand cited gap. The strongest signal we get from respondents is that anyone using AI in work should have completed a recognised, formal ethical certification. While we will not necessarily be able to teach every bit of how this technology works and how you are going to do it in that sense, if you have a workforce with really good groundings in the ethics, so they understand responsible use, the risk that they can introduce, the opportunities and the benefits, but are also critical thinkers in their use of AI, we stand a chance. That is what needs to underpin the curriculum, and to underpin skills and training, not just at school, college and university, but right the way through. We need to get our heads around that cultural change. This is not just lifelong learning as a phrase; it is literally lifelong learning, underpinned by that central ethical, responsible use for everybody. I have one final point. I cannot stress enough—and the other panel was saying it too—that contextualising the learning for the sector, the region and the role you are in is where you make it sticky. That is why we work lots with teachers at BCS. We have an AI confidence programme. It is all built on what they experience in the school and the college, because it is different from generic IT training. Maybe we do not have an affinity with it: we either are tech or are not tech and do not want to do tech training. If it is about my real job, the real children in front of me every day in the classroom and how I harness that technology for better educational outcomes, I am going to invest my time in it.

I want to move on to a debate about whether we have retraining entitlements, unemployment support and asset-building systems in the right place. We are looking at a potentially much more disrupted earnings journey for young people through to retirement. They are coming out of university with quite high levels of debt. They are having to save up more for a deposit to buy a house. They are having to put more into a pension. We are shortly to go and look at the flexicurity models in Denmark. Perhaps we could run through asset building, unemployment support or temporary unemployment support, like the Danish flexicurity model, and lifelong learning entitlements. Will, we have asked you along because we want to introduce some of the work that Nest is doing about auto-enrolment savings accounts, because that has the potential to help workers build up savings buffers in order to safeguard against disruptions to earnings.

Thanks for having me along to do that. For context, I work for an organisation called Nest Insight, which is a public benefit research and innovation centre. It is part of Nest Corporation, but with a distinct public benefit remit focused on financial security for low and moderate-income households. The issue that we have been working on addressing is the very low level of cash savings that a lot of people, particularly in low and moderate‑income households, have. One in 10 UK adults have no cash savings at all. A further 20% have less than £1,000. The solutions we have been looking at, as the Chair says, have been around payroll-deducted workplace emergency savings tools. These are tools that employers can offer their workers that enable them to save through payroll, not unlike workplace pension savings, where the money goes out before it hits your pay packet and builds up automatically in the background. The immediate focus for us has been on cash savings for emergencies, so how you help people have a bit of resilience against short-term, unexpected expenses, but we have seen that there are a number of wider potential applications of this work. That includes supporting longer-term wealth building and retirement savings behaviours, and supporting agendas to address child poverty and social mobility, which have financial resilience components to them, and violence against women and girls, where this model can help people build flee funds. We got into it through an emergency savings lens, but it has this broader suite of potential applications. In terms of how that relates to the work of this Committee and inquiry, we have already seen some ways in which these kinds of accounts can help people mitigate the labour market impacts of AI and AI-like challenges. That can include support with the cost of relocating if you need to do that to find a new job, support to bridge gaps if you have a week without earnings because you are changing jobs, and meeting up-front costs of new childcare settings or new rental if you need to move geographically to move jobs. Having this buffer fund can help labour market mobility in a number of ways. Because it was not how we got into it and we are still at relatively low levels of scale in these models, we have not looked specifically at questions around retraining. There is some evidence that these kinds of accounts could support that model. Partly, that is just a reflection of how they are designed. These accounts are flexible. They are accessible to people when they need them, for whatever purpose they might want them. There is no cap on the amount of money that can go into the accounts. Some of the employers we have worked with have expressed the idea that helping people with their financial resilience can help them be more open to retraining and seeking to progress through the firm, because it takes away the anxiety and concern of those low levels of financial resilience. We have also just seen some evidence in the US that helping people with financial resilience can reduce drop-out rates from things such as apprenticeship programmes. Retraining and then encountering a financial emergency can drive drop-out. If you have some kind of buffer fund to go back to, you see higher persistency rates within the retraining programmes. In apprenticeship programmes in the US, that has helped as well. The models that we are trialling are still relatively low scale in adoption terms. Only about 7% of UK employers offer a mechanism like this to their workers at the moment. They tend to build quite slowly because they are targeted on low and moderate-income households, so contribution rates into these accounts are not always that high. As we work to scale adoption of them and people’s savings balances build, you can certainly see a wider set of potential applications for how they could be helpful.

This could be part of the new system that we need to build to help workers acquire the resilience they might need in a more precarious, fast-moving labour market.

All our work suggests that helping people with short-term cash resilience unlocks a very wide variety of wider positive outcomes. It is also worth saying that it has direct potential benefit for the macroeconomy. Time lost to work for money-related mental ill health, and lower productivity while in work, are drains on productivity and the economy. If you can help people be a bit more financially resilient and therefore a bit more emotionally secure around their finances, and help with some of those mental health issues and other things that can undermine productivity and attendance at work, you start to build more of a macroeconomic benefit as well.

Do we think that entitlements to retraining are now in the right place in a world where, as you have just described, people are going to be genuinely having to retrain constantly for the rest of their career? We now have this thing called the lifetime learning entitlement—I think that that is what it is called. I am assuming that that is agnostic, depending on the education setting. Does that entitlement now equip workers with what they need to weather a lifetime in work in this new economy that is coming?

The lifelong learning entitlement is clearly meant to be the democratised version of student loans, if you will—so, saying everybody can get access to this to train for the equivalent of what might have been a university degree or whatever. It gives them an opportunity to potentially access funds that might well otherwise be difficult to get to. In principle, it has some really good advantages that people can draw on. For some of the folks who have been talked about by Will, it may not be appropriate to go into further debt or whatever for those sorts of things. As Julia mentioned earlier, it is getting to the stage where people genuinely feel that they have to continually skill themselves. If there is not an open and free mechanism for that universally, a mechanism that somehow funds providers and others to build the capacity that we need for that is going to be important. The LLE could well be an important part of that. Frankly, it is not well understood. As happens often with many Government entitlements, you can talk to a very significant number of employers that have never heard of it. It is coming up relatively soon and it could be a vehicle for people to take control of their own career in the way you are suggesting.

It is debt, in a sense.

It is debt, yes. It is essentially an extension to student loans.

In our recent tech priority survey, retraining was shown to be the least well-served group. While there might be quite well-understood training routes at entry level and we know there are training needs at the leadership and management end, our members are reporting that the mid-career retraining is least well served.

What does that mean?

It is not available to them.

They cannot get the retraining they need.

Some of the barriers reported by individuals are financial barriers. It suggests to me that, on Phil’s point around lack of awareness about this entitlement, you could maybe improve it by improving communications and raising organisations’ awareness of what their employees are entitled to.

The polling I have seen shows that well over half the British public think that someone else should pay for their retraining, rather than they themselves. That may mean that people are less inclined to take up more debt to fund their retraining. Do you think that there is a big enough free skills or retraining entitlement that workers have today?

I do not know whether there is an entitlement—I do not know whether others want to comment. Obviously, there is an entitlement at education level. Outside of that, and particularly in the context of AI, there is a lot of material out there; there is a huge amount available. Much of the AI boost and other such things is about saying, “Can we push it back to the technology industry and say, ‘Look, guys, if you want people to absorb your technology, you are going to have to give them the capability to get there’?”. When I used to run Cisco and we created the Cisco Networking Academy, it was on that enlightened self-interest model, in that the more people who understand networking, the more people are able to absorb networking. That is true of many of these programmes, and it has to be more like that. This is not a problem the Government can just fix. Employers can even just say that there is co-operation that is needed here.

Available training is one thing, and it being available without context is a thing. We have a plethora of AI training, but it perhaps is not singing to the people who it is trying to reach. We have evidence that, where our BCS AI training has been taken out through other trusted institutions—for example, another professional body, working with its members, with quality training—it has a much more prolific and faster uptake, and a stickier stick, basically, with people, because it means something to them. I am not sure that more stuff that is not contextualised is going to help. Training should be better contextualised, aligned to business priorities and thinking about business redesign. It should be on the job, just in time and relevant to the need. If you are retraining, it should be relevant to what you are trying to do next. That is going to be the thing that really makes the difference.

My reflection on this from the school space, thinking about young people, is that this concept of lifelong learning has a bit of a comms problem. Young people do not have a conceptual framework for what lifelong learning means. No one has really communicated that to them. There is a sense that they get some qualifications, they go on to some further study, they get a job, and that job is hopefully secure and stable. There is huge anxiety at each of those points where they have to meet those thresholds, because of the sense of the high-stakes nature of those things. Schools are working to narrate some of this—“You can have multiple careers. There are lots of options for you. You can try different things”—but it is very hard to land. If this is a concept that we want to be meaningful for everybody, young people should be leaving schools, universities and apprenticeships sensing that they have an entitlement to lifelong learning, and being able to articulate what that might mean.

Just before I bring in Dan Aldridge again, let me ask you all a very quick question: is it your view that we have to evolve either the learning and retraining entitlements that people have, or the learning and retraining systems that we have, in order to make sure that the country is well equipped for an AI future? Will, do you want to start?

For the most part, I will defer to the rest of the panel on this. Coming back to what I was saying earlier, if we think in the round about the role that financial health and financial resilience play in underpinning all these things, it is not just about the direct cost of retraining; it is about people’s ability to remain attached to the labour market as much as is possible in a world where that attachment could be very disrupted. Financial health underpins some of these other goals. Whether the savings that you build ultimately become money that you spend on retraining, or whether they enable you to retain an attachment to the labour market in other ways while those other solutions to retraining provision and the cost of retraining are put in place, it is really important either way.

It is really important to have a common spine of knowledge and skills that is continuous and does not have these big cliff edges for young people to drop off before they start the next stage, to Lauren’s point. It should also be accessible, flexible and “pick up and drop”. You should not need to make big commitments of full-time training to get to the next stage. It should be small and bite-size. Yes, some of that does require a cultural change and a comms refresh.

Julia said it earlier and she just said it again: culturally, we need to change completely. We definitely need to help with navigation. Take out all the commercial stuff; just take the navigation of the existing skills system. For most people, it is a complete mess; it is just too complex. We are trying to simplify that at the moment, but it is hard. There are 700 apprenticeship standards or whatever. There is a load of stuff to be done. We have to make it easier for people. It is all very well saying to people, “You should navigate it better,” but if you do not make it easier, people cannot find their way. You need a tube map, or equivalent, that allows people to find their way around from wherever they start to whatever they want to get to. That is so important. We have talked about this in digital skills for a while, but it is crucial for lifelong issues.

First, do not overestimate the baseline. FutureDotNow, one of our partner organisations, notes that over 60% of the UK labour workforce cannot complete the 20 essential digital tasks for work. There are over 3 million people in the UK who cannot complete any of them. There is no region where more than half the workforce are able to complete all 20. These are tasks such as online banking, changing your password and setting up an email account. These are basic, essential skills. The retail sector is the weakest. It cuts across all age ranges. It is not an older person’s issue. Half of 18 to 24-year-olds cannot do these tasks either. Those people often find themselves in SMEs. They may be lower-paid or insecure workers. It is important not to overestimate. We talk about AI a lot. We think about vibe coding and advanced tech skills, but a lot of people still have a problem with basic skills. That might be due to attitude, capability, interest or, moreover, thinking that everyone else knows more than they do and that they need someone to get on their side of the desk and do it with them. For SMEs, it needs to be bite-sized and embedded, with a low barrier to entry. Do not assume knowledge. Use CPD models that go through trusted institutions, like I described earlier. That is going to be really important to keep pace with the speed of technological change. We have to go with the SME, its business and its business priorities.

In the meeting that we were in this morning, we had people from Intuit, which does QuickBooks. It is an accounting firm, which came in and sat down with SMEs. Before the session, about 34% of them felt confident with AI. On average, 87% felt confident by the time that they had had one session. There is definitely a sense of giving people a bit of confidence to do something here. The other thing that we need to do, both on our terms and on others, is to go where the small businesses are. Talking to small businesses as a collective is hard. We need to go to their bookkeepers and accountants. We need to go to their other people. eBay was there as well, and it had a great example of a single mother who used eBay to sell a few pieces of her jewellery. So eBay gave her access to AI agents, and she has now created a bunch of agents to do her finance and other things. She has started to recruit people because it has taken the load off of her that she was worrying about before. She is now growing as a result of that. We have seen other examples where people create agents and suddenly it is the office manager who creates the best agent because they know how the system works—“You call that person, you go in there, and then you do that.” It is much more about giving them context in their world, whether it is a place that they are in or the job that they are doing.

That has been incredibly helpful. Thank you very much indeed. In the context of the first panel that we had, it is clear that we have some reforms that we have to make as a country. Thank you for helping us light that up. That concludes this panel.