Anthropic CEO Dario Amodei proposed a three-stage plan to curb frontier AI development, including external evaluators with access comparable to internal teams, government coordination among U.S. developers and verifiable international agreements. Senator Bernie Sanders introduced legislation for a 50% public stake in the largest U.S. AI companies, exercised through an independent commission. The European Union imposes systemic-risk management duties on providers of general-purpose AI models, including open-source ones, complicating enforceable pauses. The debate centers on three distinct levers: public ownership redistributing economic gains, independent access enabling inspection and legally binding halt orders imposing direct limits. A hybrid structure is proposed, combining public rules, independent evaluators and suspension powers to slow development without shifting to permanent control over general-purpose research.
Source: Read the original article
