In a bid to address escalating artificial intelligence safety concerns, Donald Trump has brokered a voluntary agreement with approximately two dozen major technology companies committing to implement AI safety controls recommended by the White House. The accord, signed on Tuesday, includes commitments to undergo independent safety audits examining cybersecurity risks, biosecurity threats, chemical hazards, and unintended AI model actions, whilst firms pledged to meet regularly and establish common safety standards. Signatories include leaders from Anthropic, OpenAI, SpaceX, Nvidia, Meta, and Google, though notably the agreement carries no legal requirement and several companies had already committed to external audits independently.

However, the arrangement has faced scepticism from observers and industry sources alike. Details remain sparse regarding how audits will function and which third parties will conduct them, leading critics to question whether the accord represents genuine progress or merely political theatre ahead of mid-term elections. Industry insiders have expressed concerns that voluntary self-policing cannot adequately address the safety problems currently affecting the sector, particularly following recent incidents at OpenAI. Trump characterised the agreement as "morally binding" despite its lack of legal force, whilst critics, including Senator Bernie Sanders, have accused the administration of prioritising industry interests over public safety.