AI Technology Regulations

Viscount Camrose Excerpts

Tuesday 30th July 2024 Lords Chamber

[Baroness Jones of Whitchurch (Lab)]

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.

[Viscount Camrose (Con)]

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?

[Baroness Jones of Whitchurch (Lab)]

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.