OpenAI, Google and Meta sign voluntary White House AI safety pact

OpenAI, Google and Meta have signed a voluntary White House AI safety pact pledging third-party audits covering hacking and biological threats. The agreement carries no penalties and no enforcement deadline, making it a non-binding commitment rather than a regulated framework.

The pact is the latest in a series of voluntary White House AI commitments as frontier labs face scrutiny over the pace of autonomous AI development.