I beg leave to ask the Question standing in my name on the Order Paper and declare my interests as set out in the register as adviser to Endava plc, the Crown Estate, Submer Ltd and Simmons & Simmons LLP.
The AI growth lab is a regulatory sandbox designed to accelerate AI innovation and adoption across the UK in a supervised and safe setting. In the recent call for evidence, we asked for views on how best to operationalise the AI growth lab. Responses are being carefully considered in ongoing policy development and will inform forthcoming legislation that we intend to bring forward.
My Lords, the AI growth lab is an excellent intervention. We have a great tradition in this country of sandboxes, and this follows on from there. It will require primary legislation, and it is not alone in that: there is an increasing number of areas where the Government say they want to act in terms of AI, all of which will require primary legislation. Rather than taking a bit by bit, Bill by Bill approach, does the Minister not agree that clarity, consistency and coherence will be better served by bringing forward a cross-sector AI Bill that would be good for the citizen, the creative, the consumer, the innovator and the investor? Will the Government take the opportunity of the upcoming King’s Speech to bring forward such a Bill?
The noble Lord has a great deal of expertise in this area and, as he knows very well, AI is a technology that has many different applications. We are committed to a context-based regulatory approach where most AI systems are regulated at the point of use. As a novel technology, it can develop in areas which cross regulatory barriers or give particular opportunities for new product and service development. That is the reason for the AI growth lab, which will model a new approach to regulation, with the power to make rapid temporary amendments to regulation to safely test and prove application. As the noble Lord says, the UK’s experience of sandboxes is emulated around the world.
My Lords, any AI growth in this country will depend on trust between tech companies and content rights holders. This will depend on robust transparency requirements being created, and quickly. However, the Government, in their report on AI and copyright last week, said they were going to sit on their hands and monitor the effects of transparency rules in other countries. Why is this country not taking the lead on this crucial issue to create a world-class, transparent copyright system for AI development?
As the noble Viscount knows, we published our report and impact assessment on AI and copyright and we have outlined three or four specific areas we will be taking forward, including digital replicas, AI labelling and looking at mechanisms for creatives to control their works online. It is clear that we need to keep considering the approach and we have not as yet found a solution that will address all the concerns.
My Lords, the Government have conspicuously rebranded their AI Safety Institute as the AI Security Institute and have been shifting their language from “safety” to “growth”. Can the Minister confirm that safety obligations will be among the regulatory red lines that can never be modified in the sandbox, and, if safety is genuinely protected, why are the Government so reluctant to bring forward safety duties in a proper AI Bill?
There will indeed be safeguards built into the AI growth lab. Modification powers would operate with robust safeguards to protect fundamental rights and safety. The lab’s design must balance the need for rapid reform and, as both noble Lords have mentioned, the importance of retaining public trust and confidence in the UK’s high regulatory standards.
My Lords, I understand the need for proportionate regulation in this area, particularly because the UK already has considerable investment and wishes to see more investment in the future. But there is one area, superintelligent AI, where there is the potential for loss of control, particularly over defence and security systems, and where some legislation is required to prevent some of these developments. In taking this forward, will the Government consider this?
The specific proposals around the AI growth lab are separate from considerations on the development of superintelligent AI and the risks that poses. We are extremely fortunate in the UK to benefit from the AI Security Institute, which is testing those models and is also able to advise the Government, and indeed the wider economy and all market participants, on the risks that superintelligent AI may pose. As we speak today, it is uncertain exactly what some of those risks are.
My Lords, we have had a plethora of initiatives on AI. We had the welcome announcement of the sovereign AI fund last week. All of them are welcome individually: Innovate UK, ARIA, the AI Security Institute, the British Business Bank, and the National Wealth Fund. Could we perhaps have a moratorium on announcements of AI initiatives and perhaps an overview and co-ordination of this plethora of bodies?
In our AI opportunities action plan we have indeed set out a plethora. That is because this is a technology that provides great potential benefits, and adoption is absolutely key. We need all areas of the economy to be thinking about the implications. We are a heavily service-based economy and it is very much in our future interests to adopt this technology and harness its benefits, whether that is in healthcare, energy efficiency or many other areas. So we will continue to focus on all areas of the development of AI, from data centres to regulation.
My Lords, my experience as Health Minister in this space is that we have fantastic innovation and fantastic pilots: the joke in the NHS is that the NHS has more pilots than British Airways. But the challenge is always the scale-up funding and, because that scale-up funding is not there, we then lose the best to America. The problem is that the AI budgets are all fragmented across hundreds of different hospitals. What are we doing to centralise those budgets so that we have the firepower to truly scale them and not lose our best British innovation to America?
Building on both noble Lords’ recent questions, we need to look across the whole of the economy. There will be a great amount of private investment going in, as well as investment from our public financial institutions. We are also thinking about how we can harness the benefit of sovereign AI here in the UK, and we are making a particular effort to think about where our sovereign AI fund is going to invest, so that the UK can benefit British frontier AI companies, not seeking total self-reliance but to build and defend comparative advantage.