Matt Clifford previously held high-level leadership positions within the UK government's core decision-making bodies and the AI Safety Institute (AISI). He was a primary architect of the UK's AI regulatory framework and played a key role in the first AI Safety Summit at Bletchley Park.
Clifford was responsible for drafting global frontier AI evaluation standards, red-teaming protocols, and government voluntary safety commitment schemes. During his tenure, he had access to classified information regarding national security review standards and undisclosed test results of frontier models.
His move to a senior role at Anthropic triggered concerns among UK MPs, business ethics review bodies, and transparency monitoring groups. Critics described the move as a 'revolving door' phenomenon that could undermine trust in the neutrality of AI regulation.
There were fears that Clifford could use non-public regulatory insights and confidential data from competitors to provide Anthropic with an unfair lobbying advantage. Following ethical scrutiny and calls for parliamentary inquiries, Clifford resigned from the company to resolve the conflict of interest.
The incident has led to calls from policy analysts and legal scholars for the UK government to tighten 'cooling-off period' rules for public officials. They argue that stricter conflict-of-interest and confidentiality regulations are needed specifically for the frontier AI sector.
