EU Technological Sovereignty Package

Baroness Gill Excerpts

Monday 20th July 2026 Lords Chamber

[Baroness Lloyd of Effra (Lab)]

The Government are taking the approach of reducing overreliance and building up our own capability. We welcome foreign investment, including in our AI growth zones, and the ability of UK companies and citizens to access the best technology available, whether it is in the UK or elsewhere. That is an important part of our being able to benefit from the developments in AI. However, I absolutely agree that AI compute is the engine behind every AI breakthrough. That is why we are investing up to £2 billion in public compute infrastructure to 2030, so that researchers, start-ups, SMEs and public services can have free access to the computing power they need to work at the frontier of AI.

[Baroness Gill (Lab)]

My Lords, the UK’s AI Security Institute does evaluate frontier models, but it lacks pre-market enforcement powers to block software releases. Pre-market enforcement would force developers to proactively prove that their models meet rigorous safety standards, such as we have for aviation, pharmaceuticals and medical devices. Are the Government open to setting a global standard for AI by backing upfront, rigorous safety evaluations?

[Baroness Lloyd of Effra (Lab)]

The AI Security Institute is one of the pre-eminent institutions in the UK in AI development and the transformations we are seeing. It works very closely with other AI security institutes. I think it was only two weeks back that it was meeting other similar institutions, developing common standards and approaches to assessment, which is one of the important ways that we keep a baseline across the globe for how we should appraise frontier models as they are developed.