White House Says Amazon, Google, Meta, Microsoft Agree to AI Safeguards

The Biden administration says it has reached a deal with big tech companies to put more guardrails around artificial intelligence, including the development of a watermarking system to help users identify AI-generated content, as part of its efforts to rein in misinformation and other risks of the rapidly growing technology. 

The White House said seven major AI companies—Amazon.com, Anthropic, Google, Inflection, Meta Platforms, Microsoft and OpenAI—are making voluntary commitments that also include testing their AI systems’ security and capabilities before their public release, investing in research on the technology’s risks to society, and facilitating external audits of vulnerabilities in their systems. 

Most of the companies declined to comment or didn’t immediately respond to a request for comment. Leaders from the companies will meet with Biden at the White House on Friday.

There aren’t enforcement mechanisms for the commitments outlined on Friday, and they largely reflect the safety practices already implemented or promised by the AI companies involved.