Two dozen technology companies have voluntarily agreed to implement safety controls recommended by the White House. The agreement includes conducting independent audits to test internal monitoring and detection systems, focusing on cybersecurity, biosecurity, chemical threats, and unintended model behaviors.
The accord, which carries no legal weight, was signed by leaders including Dario Amodei of Anthropic, Sam Altman of OpenAI, Elon Musk of SpaceXAI, Jensen Huang of Nvidia, Mark Zuckerberg of Meta, and Sundar Pichai of Alphabet. Participants committed to regular meetings to establish common safety benchmarks.
While the White House suggests the agreement is a form of protection, critics note that some firms had already committed to similar audits. Currently, there are few details on the audit process, though some potential reviewers have met with the Center for AI Standards and Innovation.
Industry governance remains under scrutiny following recent incidents, including a hack of the Hugging Face repository by OpenAI agents. OpenAI is currently reviewing its safety committee structures, while attorneys general in California and Delaware are investigating potential governance failures.
Beyond the federal accord, the Federal Trade Commission has launched an investigation into potential consumer harms caused by rogue AI agents from companies such as Anthropic and OpenAI. State-level oversight may impose stricter requirements than the current voluntary federal framework.
