AI and Creative Technologies (Communications and Digital Committee Report)

Baroness Kidron Excerpts

Friday 13th June 2025 Lords Chamber

[Baroness Kidron (CB)]

My Lords, I add my welcome to the noble Lord, Lord Massey. I also look forward to hearing from the noble Lord, Lord Evans. I thank the committee for its very valuable report, and for the determination of the noble Baroness, Lady Stowell, and her colleagues, not to split the interests of the creative, creative tech and AI communities but to try and see them as a unified whole. I really appreciated that.

It is hugely important, and the committee was right, to highlight the serious risk of the UK becoming an incubator only for foreign economies. Most importantly, the committee’s assertion that AI is not a sector but a technology is profound and so often overlooked. It is a technology whose impact on specific domains is hugely different. By failing to consider it domain by domain, we are failing to create opportunity for homegrown AI companies to dominate in specialist areas. We are also perhaps failing to consider its impact overall over the whole of society. I really did appreciate that.

My remarks are going to be about data. The report recognises that data will be central to the AI revolution. It usefully highlights both the lack of data for training and the view that more is not always more. Given that my views on the data owned by the creative industries are so liberally distributed across Hansard, I want to talk about two other datasets and make the case that it is not only cash and compute but valuing data that will support the committee’s objective of ensuring that the tech revolution is a success in the UK.

This week I received a message from someone privy to discussions about the plans for the national data library—a project that enjoys a great deal of support from Members of your Lordships’ House and that the Minister, in Committee on the Data (Use and Access) Bill, suggested made redundant the idea of designating certain valuable or sensitive datasets as sovereign data assets. I want to be clear: I believe the Minister at the time was hopeful that the issues of privacy, value to the public purse and democratic values that were integral to our concept of a sovereign data asset would be covered by the national data library scheme—same idea, different name. But my correspondent, who himself works in tech and is a serial entrepreneur, explained that currently the tech companies are heavily lobbying government to share NHS patient data in that context, on the promise of streamlining NHS bureaucracy. Sharing the entire UK population’s NHS data is an enormous decision and has privacy and cost implications, but possibly the key point for today’s debate is that it has profound implications for who, how and on what basis—and, importantly, in which jurisdiction—the benefits accrue from innovations, commercial products and services that may never have been created without the help of that dataset. This is a serious issue if patient records are shared with a UK company, but even more so if they are to be shared with a company headquartered overseas.

This bears similarities to an argument we have had for so many weeks: a concern that the Government are being primed to share something of deep personal importance to their citizens, which in this case is paid for by the public purse and in the other case belongs to other private owners, with the corresponding economic concern that British people may see little gain or, indeed, may have to meet the cost of accessing the fruits of their own data. My correspondent worried out loud that the Government were going to be hoodwinked out of a true national resource.

A similar debate is going on around the UK’s CCTV footage, which is also of enormous interest to AI companies. This trove—among the largest in the world—has many applications, one of which is the ability to model simulations on the management of large groups of people. It is undoubtedly of interest to the UK police, but once created as a product owned by a private offshore company, what is to stop it being sold to regimes across the globe? How would conflicts play out—whether in Los Angeles, Ballymena, Gaza or Congo—if those in charge had infinite scenarios from which to kettle protesters, arrest them or worse? Is that what we want to do with our precious data? If so, are there terms of engagement or is it, like the copyright debate, going to have no regulation, no powers and no upholding of UK values and laws?

More broadly, CCTV footage is some of the most valuable in the world because it shows people’s movement at vast scale. That is what is needed to train the model. If we think about it as what YouTube data is for Google, it is almost incalculably valuable. I was fascinated by the contribution from the noble Lord, Lord Willetts, and I wonder whether our data, as well as our funding, might be used to keep companies in the UK for longer.

Last week the Prime Minister confirmed that the Government have accepted all 50 of Matt Clifford’s recommendations in the AI Opportunities Action Plan. From copyright law to the data library and security issues, and even sovereign AI, discussions—in private—appear to be dominated by overseas interests. I hear constant cries from UK AI companies that they struggle to be heard, and I recall that long before the consultation on copyright was published, when I asked for a meeting with the Minister responsible for data, the Lords Minister said that he had nothing to tell me yet.

My second point is that rather than the excruciating process of missteps and ping-pong—which has not served people, Parliament or government well—if the Government had heard from a broader group of voices, or if parliamentarians had seen the draft consultation, they could have raised questions at a time that might have been more useful to all. How we share data has profound implications for our economy, our security, our national identity and even our political independence. During the passage of the data Bill, many of these issues were raised by noble Lords across the House, but the Government refused to consider them—so now we have a data Bill that looks over its shoulder rather than to the future, and the oft-promised AI Bill has been pushed away by another year. I ask the Minister in true earnestness: does she understand why parliamentarians are frustrated? We want to discuss AI in health, AI in education, AI in security and so on. Can she find a way for those in government to be more open to accepting the expertise across both Houses?

We need our data policy to benefit UK people and businesses. We need transparency from government about the deals it is making, because they all shape our economy, democracy and national identity. All the Government’s moves to improve skills, infrastructure and energy prices so that the UK AI community can thrive are extremely welcome, but on the issue of data we must have a bigger vision than offshoring the value of our data to overseas.