Liz Kendall, Lisa Nandy, Ruth Hannant and Oliver Ilott.

Q135 The Chair: We now start the second of our sessions today. The Government’s work on AI and copyright is being led jointly by the Department for Culture, Media and Sport and the Department for Science, Innovation and Technology. I am delighted that we have the Secretaries of State of both departments with us here today. Thank you so much for coming. Could the staff members who came with you introduce themselves?

Ruth Hannant: Hello, I am Ruth Hannant, the director-general for society, media and culture at the DCMS.

Oliver Ilott: I am Ollie Ilott, the director-general for AI at DSIT.

Q136 The Chair: Thank you. I do not think our Secretaries of State need much introduction, but you are welcome.

Before we turn to the main topic of today’s session, AI and copyright, I would like to comment on the recent serious issues regarding X’s AI chatbot, Grok. Speaking for all the committee, we are appalled by reports that Grok has been used to create non-consensual intimate images of women and children. This is a vile form of abuse, and I thank Secretary of State Liz Kendall for her letter to me on this yesterday, which committee members have all seen.

We welcome the Government’s announcement that they are bringing legislation into force this week to tackle this issue. We also welcome the news that Ofcom has launched a formal investigation to determine whether X has complied with its duties under the Online Safety Act. We expect it to act urgently, as I know the Secretary of State has said she expects. That is very important. We also support the Government’s intention, stated in the letter, to act on the gaps identified in our online safety legislation, such as the fact that not all chatbots are covered by the Online Safety Act.

We have some detailed questions later in the panel about transparency, which we have been working on as an issue in AI and copyright. While we have focused on that issue, mainly related to copyright, there is a view emerging that it is impossible to fully assess or mitigate the risks of harmful or illegal outputs, as we have been seeing in the case of X and Grok, without granular transparency over AI input, particularly on what data models are trained on. If there are harmful images coming out of a chatbot like Grok, how was it trained and how are those images in there? We will come on to that.

The topic of our current inquiry is AI and copyright. We read with interest the progress statement that the Government published in December, which showed that the overwhelming majority of respondents to the AI and copyright consultation wanted AI companies to pay for their training data, and only a very small minority supported any new copyright exception for AI training. In the light of this, I ask the Ministers, will you now follow the Australian Government and rule out introducing a broad new text and data-mining exception in the UK?

Liz Kendall: Thank you to the committee for inviting us. I am really glad you are looking in detail at this question; it is very important to the Government, to the country, and to me and Lisa personally. I know you will have all sorts of individual questions that you want to ask about what we are going to do, and you may get annoyed with the broad answer I give, which is that we are having a genuine reset moment. We are genuinely trying to find a way forward that seeks to back and champion our incredible, world-leading creative industry and the people who work in it, because it is so important not just for our economy, but for our country.

We are trying to find a way forward that manages to seize the potential opportunities that AI brings not just for jobs and growth, though that is vital, but for all the other huge benefits and positive ways that AI can be used. You will know that we have committed in the legislation to bringing forward a report by 18 March, which will do a whole range of things, including setting out the way forward. We are genuinely consulting and developing our proposals, so we do not have a list of areas that we can promise to you are what we are going to do and how we are going to do it, because we are genuinely consulting to try and find a way forward.

I know the issue of reward is incredibly important for creatives. So, too, is control: control of their art, what they produce. It is about controlling that and also being rewarded for it. One of the things that was crystal clear in the consultation was just how important and fundamental it is for the creative sector to have those two issues at the heart of the way forward.

Q137 The Chair: I take your point about the reset and the fact you are consulting, but it is worth saying that we have had reflected to us from the creative sector great warmth towards what the Australian Government have done in setting out that they are not going forward with a text and data-mining exception. I encourage you to keep looking at that.

Liz Kendall: We are looking in detail at what other countries are doing.

Lisa Nandy: The short answer, as Liz has said, is that we have not made a decision yet. I, for one, have really welcomed this moment. We have looked at the consultation responses, where—as you very rightly said, Chair—there was an overwhelming preference for one of those models and it was not the Government’s original preferred model.

One of the learning points for this Government has been that it was a mistake to start with a preferred model, the opt-out model. We learnt through that consultation that there are very different implications for different parts of the creative industries from the different models, and we have to take a far more nuanced approach to this and work with different parts of the creative industries to address their very serious, and in some cases, existential challenges posed by the current system, let alone any changes that we may make.

We have also learnt from looking at the experience of other countries, such as Australia, the EU and the United States, that there is no perfect solution to this, and that no country has yet managed to find an approach that does not have detrimental consequences to the two objectives that Liz set out. We believe in the potential of AI to this country and the power of our creative industries. Both will be absolutely essential to the Government’s growth mission.

There is no model out there in the world to follow that has got this completely right yet, so we are genuinely working through this, in a far richer conversation through the working groups, where instead of having two very antagonistic poles on the landscape, we have brought tech companies together with the creatives. We are finding that there is a broad consensus forming in the middle around the proper role of Government, the proper role of industry to reach its own agreements, and the things that Government absolutely needs to do, including on this issue about remuneration and fairness, that Liz talks about, but also on transparency, to try and solve this.

I appreciate, as Liz said, that it is very frustrating for the committee that we have come here without the perfect answer, but please rest assured we are taking it seriously and we know there is an urgency around it.

Q138 The Chair: It is early January now and, as Liz has touched on, the Government are committed to publishing that economic impact assessment and full report on AI and copyright in March. Will you take a definitive position on your approach to AI and copyright at that point?

Liz Kendall: One of the things we learnt last time was that coming out with a preferred option—not least not a terribly popular preferred option—was probably the wrong approach. So we have not decided that completely yet, but that is certainly one of the lessons I have learnt.

Lisa Nandy: We are keen to provide clarity around this as soon as possible. From the working groups, the number one thing that has emerged from both tech companies and creatives is that they need clarity and certainty. We want to do that but, as Liz has said previously, if we rush into this and get it wrong, we could make a mess. We are not going to rush into it; we are going to take the time to work through it with the working groups, but we appreciate the urgency around this and want to move as quickly as possible.

Q139 The Chair: What specific further evidence are you seeking that you do not already have from the working groups which restarted before Christmas and the other engagement? What further evidence is needed?

Lisa Nandy: As a first go at answering that question, lots of things have come out of the working groups, including a few key things. One relates to the issue you were talking about at the start, Chair, around Grok and the recent action the Secretary of State has taken, around insufficient protection for what we refer to as the creator’s essence, things like voice and style. We have heard some stark examples of where children’s characters have been used as part of ChatGPT to enable children to talk to those characters in some very inappropriate ways, including very sexualised conversations. It is not just a question of what that means in terms of safety; it is also a question of personality rights and other issues, which you will be aware of.

There are insufficient technical tools at the moment for some of the transparency requirements that we as a Government have committed to. One of the things we are doing is working with industry, urging them to try and find the solutions to that. The working groups have also surfaced the need for transparent processes, underpinned by legislation. That has been universally understood, especially around the issue of the use of crawlers, which poses a really serious challenge, particularly for publishing.

The other—potentially contested—thing I wanted to put in front of the committee is that there is a strong view from the working group that the Government ought to focus on creating the right conditions and frameworks, so issues like transparency, not intervening overly in licensing deals, which industry is already coming together to reach. This is essentially a demarcation of the roles where creatives and tech companies are finding their own solutions to some of those challenges, but there are certain things that they cannot do that Government ought to do. That is why we are still doing this process through the working groups, because it has surfaced that there are some areas where it needs further work.

Liz Kendall: There will be more meetings of the working groups, adding more people in and getting more information. We have done one round of those but there will be more in February.

Q140 Baroness Owen of Alderley Edge: The committee is keen to find out more about the progress that has been made to address concerns about the limitations of the existing opt-out system or rights reservation models, and to ensure that any proposed solution would be workable for both the rights holders and the AI developers of different sizes. We just heard Google say on the last panel that this option was working effectively. However, the committee has also heard evidence which says that it is a poor fit for complex copyright ownership, that they only work for rights holders who control their own domain, and that AI crawlers are simply ignoring robots.txt. I would be keen to get your perspective.

Oliver Ilott: We have had lots of conversations, both in bilateral engagement and through the technical working groups, on opt-out approaches. It is worth saying the UK does not currently have an opt-out approach because we do not have a broad exemption. Opt-outs are relevant where countries have gone down the route of creating exemption, and then the opt-out plays into that.

When it comes to the implementation of those, we have heard lots of views. One view that we have heard expressed is that the opt-out system places a regulatory burden on the rights holder, because they have to exercise that opt-out. If you are a smaller operator, such as a creative operating under your own, that puts some burden on you to go through and make sure that that is expressed properly. This might be technical; you would need some kind of interface there to help you do that.

The other thing we have heard is that if the opt-out process becomes extremely straightforward, then everyone just withdraws their material from the market. That might be fine in some sense, but if you went down the route of creating an exemption, it would be because you wanted people to access the material for training. If everyone withdraws, then you have undermined the primary policy goal you have set out there.

The other thing that we would have to bear in mind is that, when you really focus in on the granularity of how some of this material is encoded on the internet, it becomes hard. There are technical challenges in how you figure out whether an opt-out applies to this. For example, if I wrote and published a blog that was a book review, and in the course of that I was quoting from a novel, there is nothing in my blog which creates metadata that says, “This is my paragraph, this is the author’s paragraph, and this is my paragraph coming back through again”. If that author has exercised an opt-out somewhere, how do you know that it attaches to the blog that I might then have published?

We have obviously had lots of views on the idea of an opt-out as a core policy proposition in the response to the consultation but, in going through these technical working groups, we have also waded into lots of the difficulties of implementation. The EU is obviously going down this process as well and we will see how that plays out this year.

Q141 Baroness Owen of Alderley Edge: Just to push the final part of the question, do you think the proposed solution, if they went down an opt-out model, would be workable for rights holders of different sizes? I just want to really pin you on an answer on that.

Oliver Ilott: As the Secretaries of State have said, the process of figuring out how you practically deliver something would take time to work through.

Lisa Nandy: Just to add to that, the straight answer to that question is that there are challenges with the opt-out process that we had not anticipated or fully understood before we did the consultation and which we do not currently know how to surmount. It does not mean that they are insurmountable, but we do not currently have the answers to those. So, at the moment we do not have a workable opt-out proposal on the table. If you had the publishing industry here, you would hear this view expressed very clearly, underlining all of the challenges Ollie just set out with what opt-out actually means in practice.

Liz Kendall: Also, of both the large and small AI companies, lots of them did not support it either for different reasons.

Q142 Baroness Owen of Alderley Edge: What would you see as the Government’s role in any rights reservation or opt-out solution, and what do you view as a section that should just be left to industry? Where is the line between what industry does and what you anticipate that Government will do?

Liz Kendall: I go back to the point that Lisa made earlier about the role of Government in this and what only we can do, or what can be worked out within or between the different sectors. You are understandably tempting me down a road of the opt-out. I understand you asking me to set out what our role would be, but that would tend to suggest that we are backing that particular option. We know why we put that option forward, but we have also heard loud and clear from across the board why that has been rejected.

One of the things that has not come up, and certainly one of the reasons why the creative sector has not liked it, is because many would argue their permission should be sought; they should not have to opt out. From the AI perspective, we had a very strong response from people like the Startup Coalition, who said that would not work for small companies. There are a number of different concerns with it, so you will forgive me for holding fire on that in going through what our role would be in an option that has obviously got problems.

Lisa Nandy: I will just add to that a couple of things that have helped to clarify our thinking around the working groups. It is the role of industry to find technical solutions to some of these challenges, not the role of Government. We have a role in urging industry to do that, which we have been playing. But it is very difficult to enforce the rights that currently exist, let alone in a new system, and there is a responsibility on industry to find a solution to that, which we have been talking to the tech companies about. Some of that work is happening at the moment. There are some really interesting organisations out there, such as Cloudflare and others, which are finding their own solutions to this.

The second proper role for industry is around reaching licensing agreements. We have seen that the agreements and deals that have already been done are much more attuned to the nuances in their own industries. Like I said before, there are very different impacts on different parts of the creative industries.

There is a strong role for Government in introducing requirements around things like transparency, which industry cannot do itself. That cannot be done without the intervention of Government. We have committed to this as a Government; we see that as part of our proper role.

Similarly, as Liz was alluding to, a lot of the deals that have been done have worked very well for the bigger players; they do not necessarily work well for the smaller players. We are as concerned about them as we are about the bigger players, not least because the creative industries are an ecosystem—your Ed Sheerans start somewhere. We have to make sure we are protecting those people as well. There is a role for Government in stepping in where those deals are not necessarily serving the creative industries as a whole, and making sure we make this work for tech companies and creatives.

Q143 The Chair: Can I just go back to your comment about Cloudflare? We had a discussion with Google in the last session around various issues in the market. The witness we had from Cloudflare in December really called out Google and the fact that there is an imbalance in the market caused by Google having the unfair advantage they have in combining their Search with AI web crawlers. There is obviously a role in there for the CMA. It has said it will give publishers more control over their content by giving them greater choice of how their content is used by Google. However, that only deals with that tech platform.

There is a big question about what the Government could do. Could you start to take a role in addressing the fact that there is a bargaining power imbalance? It is a sort of David and Goliath situation. Publishers feel that they cannot get fair terms, and if the Goliath is Google, they have this very large advantage.

Liz Kendall: It is interesting to look at different approaches of other countries. Ollie will correct me if I am wrong, but India’s approach is of statutory licensing, where they essentially make sure that this is required then distributed, whether it is bigger or smaller. The problem with that is that it depends on there being an opt-out approach. Many creatives actually told us that was the thing they disliked most of all, because they wanted control beforehand.

Each way you come at this, there are challenges. Obviously, to govern is to choose, and that is what we intend to do. From the technical perspective Lisa is absolutely right: that is not for us to do; it is for the companies Thinking about all the different players in the ecosystem, large and small, on the AI side as well as the creative side, we have to have a system that understands that the big players come from small minnows to start with, and we need a system that works for all.

Lisa Nandy: This conversation about Google we have also had with the publishing industry in particular. When we started this process, there was a particular concern about the role that Google plays because of the search engines, so if opting out means opting out of the search engine, then you have effectively killed off your business before you have started.

What has emerged as a stronger concern since then are these AI summaries you get when you use Google. There is not a concern particularly around the AI summaries; there is a concern when the AI summaries become a competitor to the original product, using the original product to create another article that is then used by people instead of the original. That is an illustration of the challenge in this area. It is so fast-moving that new technologies are being developed all the time, throwing up new challenges which people debating this three or four years ago would not have even been able to foresee.

The challenge that Liz and I have got is that we are trying to foresee where this is going in the next three, four, five, 10 years’ time, even as technology is being invented that was not even dreamt of a few years ago. We welcome this inquiry and the work of this committee and the committees in the other place which have looked at it, because it is quite a formidable challenge to get this right.

Oliver Ilott: I will come in on some of the international comparisons the Secretary of State mentioned. Your question was about bargaining and bargaining power, and the Secretary of State mentioned India. On 8 December, India published a working paper on which they invited comments. The proposal they included in that was the idea that users would have to pay a predetermined royalty as compensation for rights holders, so you get people paying in there. That would enable lawful access. The challenge is that, in the Indian paper, that comes without an opt-out. If there was a compulsory licence royalty payment you had to make, but then people could opt out of it, you would have to negotiate with those people to get them back into it. Then, you are back into the same position of negotiating to conduct licensing.

It is worth saying that this is also an issue that has come up in Australia. Before Christmas, Australia announced that it would not go down the route of an exemption. They have set out some of the work they want to do, because there is obviously still a lot more detail for them to work through. One of the things they have said they want to look at is whether they also need a paid collective licensing framework under their copyright Act, or whether to leave it to the voluntary licensing framework they currently have. There are lots of people working through this same set of issues.

Q144 The Chair: It is worth saying that the AI overview has two issues, does it not? One is that, if you strip out the website links, you are no longer linking to the brand, so you are not going to a particular newspaper or the BBC or whatever. I was on a plane recently, looking at AI overviews, which was all I could get access to because there was no internet, and there was the whole question of, “Do I trust that answer?” There is a media literacy point here as well. If what you get is not the article from the newspaper or the BBC but a summary—

Liz Kendall: Although trusting the answers we read is not a question just for AI, is it? It is a much broader issue.

The Chair: No, but it is the action of the AI overview stripping out the links so that you cannot even go and check it for yourself. In the work we did on media literacy in the first part of last year, we were very keen to get the message across about the necessity of checking. These developments take that away.

Lisa Nandy: We recognise that. One piece of work that my department is looking at as a priority is around trust in media and whether people can trust what they are seeing and hearing. You might have heard me say it before, but we are looking as a priority at the blurring of facts and opinion, and at issues such as serving politicians presenting news programmes, for example. I see this as part of that. If people cannot see and look at the source, we are disempowering them from being able to come to informed views about whether what they are reading is accurate. So there are other challenges that we need to address.

What I do not want to suggest is that, as a Government, we see AI only as a problem, because the creative industries are increasingly at the forefront of AI and are using it to advance their work. Although there are serious challenges, which we as a Government have heard loud and clear and will not shy away from, there is also a very strong view coming from the creative industries that AI is an absolutely essential component in their future. Similarly, a really strong view is coming from companies at the forefront of the AI revolution that they need access to the great-quality human content and data that comes only from the creative industries. The two are very bound together and the futures are intrinsically intertwined. Although there are serious challenges, if we get this right, there is a great prize to be won.

Q145 Lord Knight of Weymouth: Thank you both for coming and restating your commitment to transparency, which is clearly fundamental if we are to get a regime that works and, in particular, a vibrant UK licensing market that drives the growth that we need in the economy. As the Chair alluded to earlier, we have had this issue with Grok, which you have responded to. You have to assume that the training data that will have been used for Grok included quite a lot of harmful data for it to be able to reproduce the harms that it has been reproducing. Does that create a stronger sense of urgency?

Liz Kendall: I do not think I could feel more urgent. Patience is not in any way one of my virtues and I hope that my actions over recent days have shown that. But in the Government as a whole, including in the Ministry of Justice, through the Crime and Policing Bill, we have—

Lord Knight of Weymouth: I am thinking in particular about transparency, rather than the Grok issue, which is being dealt with.

Liz Kendall: Absolutely. Just to spell it out again, we made it illegal for individuals and platforms to share or attempt to share these images. It is also now a priority offence in the Online Safety Act. That is something that my predecessor, Peter Kyle, did. We are criminalising the creation of or asking to create these images. We have said that we will bring forward the enforcement of that this week.

We are also going to criminalise the underlying model by banning nudification apps in the Crime and Policing Bill. We are coming at this to say that it is illegal. We have a very strong package of measures on it. If gaps remain, I am prepared to go further, because it is abhorrent. But it is also a matter of upholding British values and upholding the law. We should be able to uphold the law of the land in this land. I have also said previously that if we find gaps—around AI chatbots, for example, some of which are covered but some are not—we will act. I take this issue extremely seriously. You raise a really important question: where are they getting these images from? Right now, my primary goal is to protect women and children.

Lord Knight of Weymouth: In this committee, we are really proud of the work that Baroness Owen did in raising and driving this issue over the last year or two.

Liz Kendall: Totally. I am still only three or four months into this role. I know that I stand on the shoulders of many others who have fought for this for many years before.

Q146 Lord Knight of Weymouth: On transparency, Lisa, you said there were insufficient technical tools; that was one of the learnings from the workshops. We heard from Ed Conolly from Cloudflare. When I asked him specifically about more granular transparency, he said that he did not believe that this is technically complicated. How do you decide who you are listening to when you get technical advice? You have the vested interests of some big tech companies, and then you might have another big tech company, such as Cloudflare, directly contradicting. How do you make your judgment?

Lisa Nandy: That is why we set up the working groups and why we did not have just one. We wanted to cast the net very widely and hear the broadest range of voices. Although there have been some quite serious disagreements and different points of view expressed by people in the working groups, what we found is that that does not divide neatly into creatives versus tech companies. As you are alluding to, there is a debate going on within tech companies about how best to approach this.

To give you an example of some of the things that have been identified as missing by the participants in those working groups, they were essentially asking for a triple-lock approach: metadata, watermarking, and neural fingerprinting. They have identified some really significant gaps, where there will have to be more work on technical solutions. I think your question is really: why are we not getting on with legislating around transparency and—

Q147 Lord Knight of Weymouth: How close are you to a workable solution?

Lisa Nandy: To go back to the question about urgency, this was urgent when we took office 18 months ago, so there has already been a really detrimental effect on both industries as a consequence of not having clarity. We really do recognise that, as Liz said. If I might agree, patience is not one of her great virtues; it is not really one of mine either.

Lord Knight of Weymouth: We love Patience here.

Lisa Nandy: We do not want to bring forward legislation without having thought this through. All of you here who were involved in the discussions on what became the data Act will appreciate that that is not where Parliament or Government want to be. We want to work together to find a solution that we know can hold, and to legislate or regulate in order to underpin that.

Q148 Lord Knight of Weymouth: Let me attempt to quickly get a couple of principles established. First, when you have found a workable transparency solution, do you envisage it being mandatory?

Lisa Nandy: We have previously given a commitment that we will legislate for that. As Liz said, we are open-minded about being challenged on the assumptions that we have made, but that is our position; we have not yet heard a reason not to do that.

Q149 Lord Knight of Weymouth: On copyright law, previous suggestions from Government that the law is uncertain have created, as we have heard, something of a chilling effect on the licensing market. As you have suggested, universally creatives and the technical sector want certainty, so getting some certainty on licensing would be really helpful. Some witnesses have told us that they do not think the licensing laws should be changed at all.

Liz Kendall: We have heard that as well.

Lord Knight of Weymouth: Do you have a view on that?

Liz Kendall: If you look at all of these issues, including transparency and licensing, questions arise. Are there any exemptions? If so, for whom? Can they be achieved? I understand why you are rightly tempting us down there, but the truth is that this thing is a package together and has to work together. Not everybody will get everything—that is the nature of this—but we have to do a package because it is not about picking out one bit without another. We have committed to a comprehensive look, and not just on those levels; we must have something that goes across the board here. It is extremely important to have a way forward that delivers for both sides.

We have heard a lot of comments that the existing scheme works. The only caveat I would add is the point that Lisa made earlier, which is extremely important to me, too: new starters are where the next great bit of creativity will come from. I really want to make sure that we have a system that works for the small players, not just the big guys. That is a real principle for us as a Government. It has to work not just because we are a Government who believe in fairness but because we believe that that is the next new thing. We have to make sure that we protect that.

Q150 Viscount Colville of Culross: I know that you are trying to get this as a package, but I would like to ask you one specific question. We have heard again and again about the fact that having transparency around AI training data is difficult. As you have explained, some say that it is too difficult to do that granularly and some say that it is not, but what they all say is that they are very worried about the complete disclosure of the training data being a security and competition risk. Would you look at an idea where the whole training dataset was put into a black box and held by the regulator, Ofcom, and where rights holders would be able to—

Liz Kendall: Ofcom has enough on its plate right now, I think.

Viscount Colville of Culross: It could be held by the AI Security Institute, then—or somebody else—but rights holders would be able to ask whether their specific content is in there. Would that not solve the problem and ease AI companies’ concerns around competition and security?

Liz Kendall: We are always open to any suggestions. For balance’s sake in my answers, we have talked a lot about issues around transparency from the creative sector but you will know already that many AI companies, in particular the larger ones, are opposed to what they consider the onerous burdens on them in doing this. They will just carry on and do what they are doing in the States, where we do not have control.

Lisa Nandy: Yes. There is a challenge here around making sure that we are working in at least some kind of synergy with what countries are doing, as well as playing a leadership role in that, but both of us would be really keen to hear more about this; perhaps we could pick up that conversation.

As we said at the outset, this is a reset moment. We have brought people together to try to solve this issue collectively. We think that there is a strong role for the Government here, but we do not think that the Government alone can solve it. We are really keen to follow up on that conversation.

Q151 Viscount Colville of Culross: Let me take that a little further. You have raised the issue of US proprietary AI being rolled out here. What about some energy to put in a UK model, perhaps even an open-source one? We have heard about the very cheap and very effective open-source model organised by the Allen Institute for AI, Olmo 3, which apparently costs only £90 million. It also has transparency around the training decisions being made and the data being used. It seems like a wonderful model that we could use in this country; it would be great for the Government to steer something like that to happen.

Liz Kendall: What I would say, if I am allowed to answer on this bit, is that I really do want to see great British companies and great British ideas be backed. That is not just because I am a patriot; I believe in competition and want more brilliant British companies to succeed. It is healthy for this country and for the world to have a strong, proud British tech sector, based on our values, to help lead the world and provide greater competition.

I welcome inward investment from countries around the world—we are achieving that—but I also want to see something much more vibrant. We are doing a lot to back British companies in starting up, growing, expanding and succeeding, as well as in terms of ideas around how to do that and the principles on which that is based. This is a really important debate for us to have, and I want to see more of it in this country.

Lisa Nandy: We are also doing some of this ourselves. The committee will be aware of the Creative Content Exchange, which is a model that we are about to pilot. The thinking behind it is threefold.

First, the notion of fair remuneration is currently contested because there is no marketplace for the content and data that currently exists—so there is no agreed price. We think that, by creating a marketplace for that content and data, we can resolve this issue.

Secondly, there is a genuine challenge around enforcement. At present, most of the disputes around the use of content and data are settled through the courts. That inevitably disadvantages the smaller players, who are, as Liz said, a point of principle for this Government. We think that the Creative Content Exchange will help deal with that challenge of enforcement, particularly for smaller players.

Thirdly, we think that there is a huge opportunity here. In speaking to tech companies that are developing and refining their AI models, the conversation used to be very much about the quantum of data, but it is now about the quality of data. The UK is a real leader because of the strength of our creative industries in high-quality content and data. We think that opening that up to tech companies in a fair and legal way will bring them huge benefits and attract a lot of companies from all over the world to do business with us.

Also, think about some of the major institutions in this country. Whether it is the National Archives, the British Library or the BBC, they are sitting on a goldmine of content and data that is simply not being used at the moment. We think that there is enormous potential here, as well as a role for the Government to help—that is, not run the Creative Content Exchange but act as a catalyst to bring it into being. That could be world-leading if we get it right.

This idea of pushing at the boundaries of what is possible—and not just learning from the experiences of other countries but leading—is absolutely at the centre of what we are trying to do.

The Chair: That segues very nicely into the question from Lord McNally.

Lord McNally: I was very impressed at the paling from both of you when you saw that Jim Knight was on our committee.

Lisa Nandy: That is just how we react to Jim.

Q152 Lord McNally: I want to take up the point about the Creative Content Exchange. Before I do, let me offer a little reminiscence. I have been in the Lords for 30 years. When I first came in, one of the personalities in the Lords was Ted Willis—Lord Willis—who was famous for being the inventor of “Dixon of Dock Green”. I always remember him once saying of a Minister, “His problem was that he always carried the imprint of the last arse that sat on him”—probably an unparliamentary intervention. But it was very welcome last night, Liz, to watch you making the Statement and answering questions. It was not exactly a breath of fresh air, but it was nice to hear a government Statement that showed a genuine outrage and proposed some solutions to the problem.

I think the problem that that you two face now is that it is polarised between a creative industries sector of which we are very proud, which creates lots of stuff, and so on, and a desire to do what you just said about promoting British industry more generally.

Just to put it into one thing, earlier this afternoon we were talking to one of the big companies. Of course, we can question them hard about their transparency and accountability, but the witness threw in that they are going to invest £5 billion in the British economy, and that always carries weight—there is a question at the end of this, by the way.

One of the earlier witnesses also said—I am not exactly using his words—that you say you are taking a rain check, et cetera, but we have almost pulled into a layby and the rest of the world is passing by. We are talking about a change that is not in years or decades, but months. I think one of you mentioned two or three years of really fundamental changes. So how do we keep up to speed on the broadest thing?

Now we come to the question. You have just given great hope in describing this Creative Content Exchange as a trusted marketplace to support licensing and access to digitalised cultural assets for AI training. Is this because you feel that the present system that we have will just not deliver that kind of co-operation, and if so, how quickly will this be up and running?

Lisa Nandy: In a nutshell, we think that we need more creative solutions in this space. We look at the other countries that have taken action, and they have run into their own difficulties with the approaches that they have taken, so we think we need to bring a fresh approach. But I think, Ruth, you could talk in more detail.

Ruth Hannant: Absolutely. The content exchange is absolutely not a replacement for existing marketplaces that are popping up—we know that there are quite a few different marketplace ideas around there—and it is also not a replacement for existing deals. As the Secretary of State outlined, there are three problems that we see. I think one is data access, in particular to really high-quality data; AI firms really need that to be able to grow. Then there is that really important thing about unrealised economic value. Actually, there is quite a lot of value in our cultural and creative assets which is currently untapped, and there is an awful lot of data which is either not available for AI developers at the moment or which is not even digitised at the moment. So, one of the things that we are trying to do through the content exchange is to answer that question of: “We think there’s lots of value in these things. Is there?” So it is a research undertaking. It is being funded by the UKRI through the R&D Missions Accelerator programme, and we are testing these theses.

The questions that we are trying to look at are: is there commercial value in the sort of data that a lot of our public bodies are holding? Can you achieve a fair value exchange between those data providers and the market and can we then build that to scale? One of the other things we are keen to look at is whether there are agglomeration benefits that you could get by bringing lots of different types of data together into something like the exchange. The final thing is, if we have proved all this, then, as the Secretary of State was saying earlier, it potentially provides a mechanism to enable smaller providers and, importantly, individuals to get access to market through a really trusted provider.

That is what we are trying to look at, but it is a research undertaking, and we are going to launch a pilot later this month that will run for 12 months, with a review point in the middle of that, to see whether we think this is an idea that works, because we are trying something new and we are testing and learning. If that review point works, we will look at what its governance structures should be and how we would set it up, and it will then evolve into something. As the Secretary of State said, we are trying to incubate something quite innovative.

Q153 Lord McNally: That is very helpful. Are you sure that you are not treading on toes or duplicating things that have already been done?

Ruth Hannant: Absolutely. What we are looking to do is to build on approaches that we have seen, but to answer those quite specific questions—in particular, looking at whether there is value in public assets that we can commercialise which has not been commercialised before, which then potentially has spillover benefits to helping bits of the creative industries that have not also been able to access these sorts of deals.

Lisa Nandy: It absolutely goes back to that question at the start around what the proper role for Government is in this and what is the proper role for industry. There are quite a few public institutions that are sitting on what we think is a goldmine of content and data that could be enormously beneficial to AI companies. There is not a level of trust there at the moment about sharing and opening that up. There also is not an incentive to do so, because essentially it costs money to digitise and to order that data and content. If you are an organisation that has not had a huge amount of funding over the last decade and a half, you would be reluctant to make a priority, but you might be incentivised to make a priority if you could see that there was value in it. There is a proper role for Government here because, frankly, nobody else is able to do what Government can do in this space, acting, as Ruth said, as a catalyst and an incubator, not as the solution to this but as the way of helping to bring that solution into being.

Lord McNally: One advantage of being in the House of Lords is that you get the benefit of history. Some 15 years ago, I was on a Joint Committee with Francis Maude in the early days of the coalition, which was looking at releasing the NHS database for the benefit of mankind, and we were sunk below the waterline by the Daily Mail, which told its readers that we were about to release everybody’s health records to the public. But it is still there, and you are right that it is not just the NHS. We have some amazing assets there, if properly managed, and I am glad my recall of history is being supported by you.

The Chair: Jim is anxious to have a last word.

Q154 Lord Knight of Weymouth: This is a question for Liz. I think that as a committee we all share your passion around driving growth in the economy through AI and allowing that to develop. We as a committee published a report on growing AI and createch businesses here a year or so ago. Essentially this was an issue, but there were really big issues around access to compute, to talent, to energy, et cetera. Do you have a sense in your prioritisation of things of where this sits in terms of being able to maintain, attract and grow an active AI?

Liz Kendall: I really do want to focus on this and I also want to say something. AI is becoming the engine of economic power and hard power—the power of nation states. I want to make sure with my colleagues that Britain wins the right race, that we make sure our incredible universities, researchers, brilliant start-ups, creative sector, life sciences, defence and financial services seize this, but that we also understand Britain’s role in the world in this. We have to be really realistic about the degree of economic and hard power that AI brings. We have a role, a duty and a responsibility in this country to seize the benefits, but to shape it to work for all based on our values.

Without growth, we do not tackle living standards, and we have no money to back our vital public services. There is no path to better growth in this country without AI, science and tech. I know you are all realists in here. That is the truth of it. We face huge geopolitical changes here. Britain’s role in this is extremely important. I am very ambitious for this country. I believe we can use this to develop jobs and growth in every part of the country. We are opening AI growth zones in the north-east, in north Wales, in south Wales, and there will be other ones coming, bringing jobs and growth everywhere.

The main things people talk to me about is that we need the compute, we need the talent, we need to keep our talent here, and we need greater competition in the world. We are part of that, and we need to seize these benefits. I want to do this in a way that supports our incredible creative sector. If you could combine AI with our world-leading strengths—life sciences, the creative sector, defence, financial services—that is the race we win, but we have to do it based on our values.

I want to solve this issue, as does Lisa, but we cannot lose this moment as a country to deliver for our people and to shape this technology, which will shape us whether we like it or not. I want it to be for the good of people in every part of the country and based on our values. I feel very strongly that to be in this job in this country at this time is really important, because we can both win and shape this for the future of our people and the future of the world. AI copyright is one extremely important part of a whole series of things that we need to do based on our values, and we are determined to deliver.

Q155 The Chair: We have reached the end of our time, sadly. There were an awful lot more questions we could have raised, but it was heartening to see you both today and to hear what you covered with us. We have spent a number of months on this inquiry and our team now has to write up the report, which is always the hard bit. I take heart particularly from you talking still about the reset, which is important. We cannot underestimate the chilling effect that a number of the missteps that happened before had on developments. We have touched on those here.

The signals going from Ministers to the outside world are very important, and you are still consulting, but we need solutions. People are starting to say that it is really good to be consulting, but we have had a lot of consultation—years and years of it. I understand why you cannot give us the exact dates of when we are going to see solutions, but we want them.

The points we made about transparency and its enforcement are very important parts of finding a solution. That has particularly come alive with the knowledge of X’s Grok and the fact that it must have trained on pornography and all kinds of horrors. It is not just a question of transparency opening up the remuneration solutions, but that it will potentially keep children and women safe. Licensing—certainly missteps and things said in the past—made the difference, and we need support to see licensing markets opening up.

Thank you very much. I suppose we could just say: go carefully through these things but try and move to solutions that work for us.

Liz Kendall: Carefully but quickly and definitely.

Lisa Nandy: That is our mantra.