Center for American Progress

The analysis urges the United States and China to work toward five goals: Establishing an “AI red phone” for urgent safety concerns. Defining “genuine security threat” from frontier AI that would impel steps to pace the frontier in the future. Announ...

Physicians should not bear sole liability for harms arising from such model-level risks. AI developers should also be held liable for harms from AI due to their knowledge about the design of those AI tools.