MPs want a single regulator with power to block AI models in the UK
Key Points
- Joint Committee on Human Rights calls for a single, independent AI regulator on a statutory basis
- Regulator would test AI systems before release and block those posing unacceptable risks
- Powers would include codes of practice, transparency requirements and sanctions
- AI Security Institute currently works with developers on a voluntary basis only
- Committee says no country has an AI regulatory approach that is fit for purpose
MPs and peers on the Joint Committee on Human Rights want a single statutory AI regulator with the power to test models before release and stop them reaching the public.
In a new report on human rights and the regulation of AI, the committee said no country, including the UK, currently has a legislative and regulatory approach to AI that is fit for purpose.
It called for a new AI Bill to create a single, independent oversight body on a statutory footing. The body would act as the central point of contact for concerns about AI use and monitor AI harms and risks across the UK.
The committee said the regulator should have strong powers to hold AI providers to account, including setting codes of practice, imposing transparency requirements and sanctioning wrongdoing.
“A single AI regulator should be established to set policy, monitor performance and with the teeth to ensure enforcement,” said Alex Sobel, chair of the Joint Committee on Human Rights.
UK regulators currently lack the power to test and evaluate AI systems before companies release them, the committee said. They also cannot prevent a release where they believe a system poses unacceptable risks.
Model developers work with the AI Security Institute on a voluntary basis, and the institute has no statutory power to compel them to do anything.
The report described the existing legal framework for AI as fragmented and difficult to navigate. It said UK laws that could apply to AI mostly do so at the point of deployment, which leaves responsibility with users of the technology rather than the companies that design it.
The committee said that harm-specific and sector-specific legislation leaves gaps in protection and fails to address preventable human rights risks.
It also warned that AI threatens the right of individuals to seek an effective remedy against decisions made about them. A lack of transparency over how organisations use AI in decision making, and a lack of effective ways to challenge those decisions, puts people in the UK at risk, the committee said.
“Fundamentally, this is about making sure that you, as an individual, know when AI is being used in the decisions that affect you,” said Sobel.
The committee said the UK should continue to play a leading role in international AI summits and sign up to international declarations and agreements on AI safety.
