Trump Signs Artificial Intelligence Safety Pact With Top Technology Leaders

U.S. President Donald Trump has published a one-page artificial intelligence accord on his Truth Social platform following a meeting with major technology chiefs in Washington. The document, which the administration describes as morally binding, bears the signatures of Mr. Trump alongside the heads of Google, Anthropic, Meta, OpenAI, the X social media platform, and NVIDIA.

Accord Controls

The agreement outlines four distinct layers of controls for artificial intelligence companies. First, it urges firms to partner with external evaluators to conduct independent reviews of their safety systems. Second, it calls for an independent board committee to oversee these reports from both internal and external teams. Third, companies must use internal controls during model training and deployment to guard against cybersecurity, biosecurity, and chemical risks. Finally, firms need an internal team to verify that all detection and monitoring tools function correctly.

Future Outlook

The White House document suggests that these voluntary steps could eventually become codified into formal laws and regulations. Participating companies have agreed to meet on a regular basis to set safety standards and share best practices.

Context

Governments and technology firms worldwide are increasingly debating how to regulate rapid advancements in artificial intelligence to prevent potential risks. Such voluntary agreements often serve as a starting point before official legislation is drafted and passed into law.