Technology and AI
A research agency chair joins a model developer and keeps the chair
By Staff Writer | 3 September 2026

The man who wrote the government's artificial intelligence plan has taken a job at one of the companies it benefits, and intends to stay on as chair of the agency that hands out public research money. The Commons committee that watches both says the arrangement is within the rules and still creates a conflict.
Matt Clifford has joined Anthropic as managing director of international affairs, leading its work with governments outside North America. He has said he will remain chair of the Advanced Research and Invention Agency and keep his role at the venture firm he co-founded, while stepping back from everything else.
The reason that combination drew a statement from a select committee within a day is the route he took to it. He was commissioned in July 2024 to write the AI Opportunities Action Plan. The government published it in January 2025, adopted all fifty of its recommendations, and appointed him as its AI Opportunities Adviser. He has chaired the research agency, which invests public money in high risk research, since 2022.
What the committee said
The chair of the Commons Science, Innovation and Technology Committee published a statement on 2 September. It does not dispute that the move is permitted. It says he has made a contribution to the development of UK artificial intelligence policy, that it is a matter of some concern that he is taking his knowledge, experience and government contacts to a leading global AI company, that this is entirely within the rules, and that the committee would not want to discourage movement between the public and private sectors.
Then it turns. His intention to remain as chair of the agency, which invests taxpayer money in research much of it related to or enabled by artificial intelligence, creates what the statement calls a clear conflict of interest.
The committee looks forward to the AI Minister setting out robust safeguards and clear boundaries to ensure confidence in ARIA's independence.
Dame Chi Onwurah MP, Chair of the House of Commons Science, Innovation and Technology Committee
The committee has not asked him to resign either post. It has asked a minister to publish the terms on which he holds both.
Recusal is the answer, and recusal is the question
The reported arrangement is that he will recuse himself from matters involving his new employer at the agency. That is the standard answer and on its own terms it is a reasonable one. It also has a known weakness, which is that a chair does not only decide the items in front of the board. A chair shapes which items reach the board, what the agency chooses to fund, which fields it opens and which it closes, and none of that is captured by stepping out of a room when one company's name appears on a paper.
Nothing in the committee's statement alleges that anything improper has happened. Its request is narrower and harder to refuse: publish what the boundaries are, so that confidence in the agency's independence does not rest on an undertaking nobody outside has seen.
Why anyone outside Whitehall should care
The pattern is not confined to artificial intelligence and it is not new. A person writes the strategy, the strategy is adopted whole, the person then chairs the body that allocates money under it, and then joins a company that stands to receive some of that money. Each step is defensible. The sequence is what a competitor bidding into the same programmes has to weigh.
For anyone who bids for public research funding, or contracts with a body that does, the practical question is whether the agency's declarations of interest and its recusal arrangements are published, and whether they can be inspected before a competition rather than after it. That is the document the committee has asked for. Whether the minister produces it, and how specific it is, is the thing to watch.