Question to the Department for Science, Innovation & Technology:

To ask the Secretary of State for Science, Innovation and Technology, what (a) short, (b) medium (c) and long-term actions he is taking to help anticipate and mitigate the potential risks of AI loss-of-control.

AI models have the potential to pose novel risks by behaving in unintended or unforeseen ways. The possibility that this behaviour could lead to loss of control over advanced AI systems is taken seriously by many experts.

The AI Security Institute (AISI) is researching the development of AI capabilities that could contribute towards AI’s ability to evade human control, as well the propensity of models to engage in misaligned actions.

Furthermore, through the Alignment Project – a funding consortium distributing up to £27m for research projects – AISI is supporting further foundational research into methods to develop AI systems that operate according to our goals, without unintended or harmful behaviours.

The Government has been clear that we will legislate on AI where needed but we will do so on the basis of evidence where any serious gaps exist.