President Donald Trump and leaders from six major technology companies signed a voluntary agreement on Tuesday to oversee AI safety. The signatories include Dario Amodei from Anthropic, Sundar Pichai from Google, Mark Zuckerberg from Meta, Greg Brockman from OpenAI, Jensen Huang from Nvidia, and Elon Musk from xAI.
The accord mandates a four-layer oversight structure. Companies must implement internal controls to monitor model behavior during training and release, specifically addressing biological, chemical, and hacking risks. These controls require verification by internal teams and independent third-party auditors.
Findings from internal teams and external auditors will be reviewed by an independent board committee. Participating firms are expected to meet regularly to define safety standards. Trump described the commitment as morally binding, noting that the agreement currently lacks legal enforcement.
The administration plans to appoint an overseer in the near future and form a committee consisting of approximately 10 members. While the document suggests that codifying these measures into law may be considered later, Trump has rejected calls for immediate federal regulations, citing the need to maintain a competitive advantage over China.
