Open Artificial Intelligence Service

Earl of Devon Excerpts

Wednesday 3rd September 2025 Lords Chamber

[Lord Vallance of Balham (Lab)]

The AI Security Institute was set up to look at the potential risk of new models. It works closely with model developers and gets access to models early, tests them for things that might be problematic, and is an important source of identifying possible issues. There are of course much broader questions, such as the one that has just been asked, which are beyond what happens inside government. That is why the work the Alan Turing Institute started, looking at some of those issues, is important. I am very pleased that many of the people who led that are now being established in academic positions and will continue to address these very important questions as we go forward.

[The Earl of Devon (CB)]

My Lords, we have heard repeatedly about the importance of data to the success of our AI industries. We are also well aware that people’s relationship with AI is very much guided by the fear of their data being misused. I understand that Sir Tim Berners-Lee is looking closely at this issue and how we return the ownership of our data to us as individuals from companies that are harvesting it and using it without our consent. What are the Government doing to pursue this and to look at how we as citizens can control our own data in this new AI world?

[Lord Vallance of Balham (Lab)]

Of course, this is a major issue, particularly in health data. I was intrigued when speaking recently to colleagues in Denmark. They made the point that they have a very simple message: they provide a health service free of charge, and in return citizens of Denmark are expected to provide their data to improve that service, but they still own the data. The question of how we manage that with data ownership in the UK, in health and beyond, is one of the things the National Data Library and Health Data Research UK will have uppermost in their minds as they develop their services.