President Donald Trump introduced a one-page artificial intelligence accord on Tuesday. The document, which Trump described as morally binding, was signed by the leaders of several major technology companies, including Google, Meta, OpenAI, Anthropic, Nvidia, and X.
The agreement outlines four voluntary measures for participating firms. Companies are expected to implement internal controls to monitor models for cybersecurity, biosecurity, and chemical threats, while ensuring systems do not access technical infrastructure in unintended ways.
The pact requires companies to establish internal teams to verify these controls, engage external auditors for independent assessments, and form independent board committees for oversight. The document suggests that these voluntary steps could potentially be codified into law in the future.
The accord includes no penalties for non-compliance. Trump announced plans to appoint an official to lead the initiative and form a committee consisting of approximately 10 people to oversee the effort. The administration maintains a preference for self-policing over strict regulation.
