29 January 2024
The White House has announced a number of major policy developments for AI in the US, with significant implications for tech companies.
Following Executive Order 14110 in November, the Biden administration today announced a number of major developments to promote AI safety and US national security.
Key developments include:
🔹 Using the Defense Production Act to compel developers of the most powerful AI systems to report ‘vital information’ to the Department of Commerce
◾ This includes reporting AI safety test results and the existence of large computing clusters able to train these systems
🔹 Proposing a draft rule to compel US cloud companies providing cloud services for training AI to non-US organisations to report this to the Department of Commerce
◾ If finalised, the proposal would require cloud providers to alert the US government when overseas clients train the most powerful AI models using their cloud infrastructure
🔹 Receiving completed AI risk assessments for critical infrastructure from nine US government agencies as the basis for future federal action on AI safety
◾ These included risk assessments from the Departments of Defense, Treasury, and Health and Human Services being submitted to the Department of Homeland Security
Deputy Chief of Staff Bruce Reed will convene the White House AI Council today to discuss completion of the 90-day action plan in the Executive Order. The AI Council consists of top officials from a broad range of federal departments and agencies.
Details on how the information-sharing requirements will work in practice will emerge in the coming weeks. The US Federal Government appears to be taking control of AI safety, as a national security priority.
How this will affect individual tech companies, and discussions on global AI standards with the UN Secretary-General's AI Advisory Body and future AI Safety Summit meetings remains to be seen...