Matt Clifford, a former adviser to Sir Keir Starmer and Rishi Sunak, has joined Anthropic as managing director of international affairs, prompting concerns over his decision to remain chair of the UK’s taxpayer-funded Advanced Research and Invention Agency (Aria).
Clifford announced the appointment on X on Wednesday, saying he would lead Anthropic’s work with governments outside North America. He said he would be based in London and work with governments across Europe and Asia-Pacific, arguing that decisions shaping AI’s impact should involve “industry, governments and civil society working together”.
Anthropic said Clifford’s role will involve “leading the company’s engagement with governments” across international markets. Clifford said people should have “real agency” over how AI is developed and used in their societies, and that Anthropic took this principle seriously.
Clifford previously served as an unpaid AI adviser to Sunak and helped organise the 2023 AI Safety Summit. He later advised Starmer and authored Labour’s AI Opportunities Action Plan, which called for changes to copyright law to align with EU rules.
The appointment has drawn criticism from Chi Onwurah, chair of the House of Commons Science, Innovation and Technology Committee. She said Clifford’s decision to remain at Aria created a “clear conflict of interest”, given the agency spends taxpayer money on technology and research projects, including work involving AI.
Clifford said on X that he would remain chair of Aria and Entrepreneurs First, the technology venture firm he co-founded, while stepping back from his other activities to focus on his Anthropic role.
Anthropic said Clifford would recuse himself from any Aria matter involving the company. The company added that its commercial relationship with Aria was managed by a separate team and that Clifford had no role in procurement or programme-level spending decisions.
The UK government said it had agreed safeguards with Aria, including Clifford’s recusal from business that could create a conflict of interest. A government spokesman said the arrangements were consistent with established rules for public officials and would remain under review.
Clifford has previously argued that AI companies should be allowed to use copyrighted material for training unless rights holders explicitly opt out. The position has faced opposition from publishers and creative industries, while the government has delayed proposed changes to copyright law.
The appointment comes as Anthropic faces legal action from music publishers Sony and Warner over allegations that it used tens of thousands of songs without permission to train its AI systems. Anthropic has denied the allegations and said it would defend itself in court, while it agreed last year to pay $1.5 billion to authors who alleged that their works had been used without permission.
