AI Technology Regulations

Question

Asked by

To ask His Majesty’s Government what plans they have to regulate artificial intelligence technologies.

My Lords, as set out in the King’s Speech, we will establish legislation to ensure the safe development of AI models by introducing targeted requirements on a handful of companies developing the most powerful AI systems. The legislation will also place the AI Safety Institute on a statutory footing, providing it with a permanent remit to enhance the safety of AI. We will consult publicly on the details of the proposals before bringing forward legislation.

I thank my noble friend the Minister for her reply and congratulate her on her appointment. There is no doubt that AI will be an important part of the economic growth that is this Government’s priority, but there are also growing concerns about the potential harms being caused by this technology, in particular around the creation of deepfake content to pervert the outcome of elections. What is the Government’s view on that potential harm to democracy, and are there any plans to extend the regulation to political advertising, as recommended in the 2020 report to this House from the Democracy and Digital Technologies Select Committee?

I thank my noble friend for those good wishes. Of course, he is raising a really important issue of great concern to all of us. During the last election, we felt that the Government were well prepared to ensure the democratic integrity of our UK elections. We did have robust systems in place to protect against interference, through the Defending Democracy Taskforce and the Joint Election and Security Preparedness unit. We continue to work with the Home Office and the security services to assess the impact of that work. Going forward, the Online Safety Act goes further by putting new requirements on social media platforms to swiftly remove illegal misinformation and disinformation, including where it is AI-generated, as soon as it becomes available. We are still assessing the need for further legislation in the light of the latest intelligence, but I assure my noble friend that we take this issue extremely seriously. It affects the future of our democratic process, which I know is vital to all of us.

My Lords, I welcome the creation of an AI opportunities plan, announced by the Government, but, as the noble Lord, Lord Knight, says, we must also tackle the risks. In other jurisdictions across the world, including the EU, AI-driven live facial recognition technology is considered to seriously infringe the right to privacy and have issues with accuracy and bias, and is being banned or restricted for both law enforcement and business use. Will the Government, in their planned AI legislation, provide equivalent safeguards for UK citizens and ensure their trust in new technology?

I thank the noble Lord for that question and for all the work he has done on the AI issue, including his new book, which I am sure is essential reading over the summer for everybody. I should say that several noble Lords in this Chamber have written books on AI, so noble Lords might want to consider that for their holiday reading.

The noble Lord will know that the use and regulation of live facial recognition is for each country to decide. We already have some regulations about it, but it is already governed by data protection, equality and human rights legislation, supplemented by specific police guidance. It is absolutely vital that its use is only when it is necessary, proportionate and fair. We will continue to look at the legislation and at whether privacy is being sufficiently protected. That is an issue that will come forward when the future legislation is being prepared.

My Lords, would the Minister agree that the way to regulate AI is principles-based, outcomes-focused and input-understood, and always, where appropriate, remunerated? To that end, what is the Government’s plan to support our creative industries—the musicians, writers and artists who make such a contribution to our economy, society and well-being, and whose IP and copyright are currently being swallowed up by gen AI, with no respect, no consent and no remuneration? Surely it is time to legislate.

The noble Lord raises a really important point here and again I acknowledge his expertise on this issue. It is a complex and challenging area and we understand the importance of it. I can assure the noble Lord that it remains a priority for this Government and that we are determined to make meaningful progress in this area. We believe in both human-centred creativity and the potential of AI to open new creative frontiers. Finding the right balance between innovation and protection for those creators and for the ongoing viability of the creative industries will require thoughtful engagement and consultation. That is one of the things we will do when we consult on the new legislation.

My Lords, artificial intelligence poses a risk not only to high-profile issues such as existential threats and safety, but also potentially to public standards—a matter on which the new Government have made many statements. Areas such as objectivity and accountability are potentially undermined through the use of AI for official decision-making. Can the Minister confirm that those aspects of the risk posed by AI will also be properly considered as steps are taken to move towards regulation?

The noble Lord is right that there are issues around the risks in the way he has spelled out. There are still problems around the risks to accuracy of some AI systems. We are determined to push forward to protect people from those risks, while recognising the enormous benefits that there are from introducing AI. The noble Lord will know I am sure that it has a number of positive benefits in areas such as the health service, diagnosing patients more quickly—for example, AI can detect up to 13% more breast cancers than humans can. So there are huge advantages, but we must make sure that whatever systems are in place are properly regulated and that the risks are factored into that. Again, that will be an issue we will debate in more detail when the draft legislation comes before us.

My Lords, let me start by warmly welcoming the Minister to her new, richly deserved Front-Bench post. I know that she will find the job fascinating. I suspect she will find it rather demanding as well, but I look forward to working with her.

I have noted with great interest the Government’s argument that more AI-specific regulation will encourage more investment in AI in the country. That would be most welcome, but what do the Government make of the enormous difference between AI investment to date in the UK versus in the countries of the European Union subject to the AI Act? In the same vein, what do the Government make of Meta’s announcement last week that it is pausing some of its AI training activities because of the cumbersome and not always very clear regulation that is part of the AI Act?

Again, I thank the noble Viscount for his good wishes and welcome him to his new role. He is right to raise the comparison and, while the EU has introduced comprehensive legislation, we instead want to bring forward highly targeted legislation that focuses on the safety risks posed by the most powerful models. We are of course committed to working closely with the EU on AI and we believe that co-ordinating with international partners —the EU, the US and other global allies—is critical to making sure that these measures are effective.

My Lords, I also express my good wishes to the Minister and say that my noble friend Lord Knight has raised an exceptionally timely Question on what is, increasingly, a major challenge for the UK: AI. I was pleased to work with my noble friend recently—or at least a few years ago now—on the Future of Work Commission. My area of concern is work. Can the Minister expand further on the use of regulation and the timeline, if possible? Does she have concerns about the potential loss of employment, despite the many opportunities?

I thank my noble friend for his question. He is right that there are huge opportunities from applications of AI in the workplace, but also a number of areas of cause for concern. As he knows, there have been very worrying cases where people have been sacked by a computer, sometimes incorrectly. We want to make sure that that is not possible in future and that people have more rights to be dealt with by a human being rather than by a machine.

This was an issue that came up for a great deal of debate in the last data protection Bill, which did not make it through the wash-up, but the new smart data and digital information Bill, announced in the King’s Speech, will hopefully pick up some of those issues and we will look at how we can ensure that workers are protected.