Matt Clifford resigns as ARIA chair over Anthropic conflict of interest

Matt Clifford has resigned as the chair of the Advanced Research and Invention Agency (ARIA), the UK government's high-risk research unit. The departure follows Clifford's acceptance of a full-time position at Anthropic, a San Francisco-based AI company, sparking a backlash over governance and impartiality.

Clifford's new role at Anthropic involves leading the firm's engagement with governments outside the United States, including the UK. While he initially attempted to maintain both positions through recusal agreements, the overlap drew sharp criticism from senior policymakers. Chi Onwurah, the Labour chair of the Commons science, innovation and technology committee, described the arrangement as a "clear conflict of interest."

In a statement, Clifford said, "I have decided to step down to ensure my new role at Anthropic doesn’t become a distraction from Aria’s incredible work." Despite the resignation, Clifford will remain in his post at ARIA until November 6, 2026, with specific safeguards in place to manage the transition.

A central figure in AI policy

Clifford has been a cornerstone of the UK's strategy to lead in artificial intelligence. He served as an AI adviser to Prime Minister Keir Starmer in early 2025 and played a pivotal role in the 2023 AI safety summit. His leadership of ARIA was intended to drive "moonshot" science—funding frontier research that is too risky for traditional grant bodies but possesses the potential for transformative breakthroughs.

This incident occurs amid a growing trend of "revolving door" appointments between Whitehall and the private sector. Other high-profile former officials, including Rishi Sunak and George Osborne, have similarly transitioned into roles at AI giants such as Anthropic, Microsoft, and OpenAI.

Implications for governance

The controversy underscores a deepening tension between the UK's ambition to be a global AI superpower and the necessity of impartial state governance. Critics argue that when the architects of national policy move directly into the firms they once helped regulate or advise, it risks compromising the government's ability to maintain a precautionary approach to AI safety.

Beeban Kidron, a crossbench peer, emphasized the need for transparency to maintain public confidence, stating, "If we are to rebuild trust in the government, we need to see a clear line between tech interests and those of citizens and the nation."

What's next

Attention now turns to how ARIA will replace its chair and whether the government will implement stricter rules regarding the movement of officials into the AI industry. While Clifford's exit resolves the immediate conflict, the broader debate over the independence of state-funded research agencies remains open as the UK continues to navigate its relationship with the dominant US-based AI firms.