Dario Amodei’s AI Policy Agenda Calls for Pacing Frontier Development

Dario Amodei, co-founder of Anthropic, has outlined a policy agenda to pace the rapid development of frontier AI systems, which are outpacing public policy responses. In his essay, "Policy on the AI Exponential," Amodei proposes a five-area action plan to regulate the development and deployment of advanced AI systems while preserving their potential benefits.

The five policy areas include frontier-model testing, employment effects, civil liberties, management of downstream technologies, and international coordination among democracies. The essay emphasizes the gap between accelerating AI capabilities and institutions designed for slower-moving technological change.

While the essay does not call for a complete halt to the use of AI, it suggests that rules, safeguards, and economic conditions surrounding powerful models may change rapidly as policymakers respond to faster technical progress. The proposed FAA-like approach for mandatory third-party testing of frontier models and government authority to block unsafe deployments could create a more formal process for evaluating the most capable AI systems before deployment.

The essay also calls for better data on AI-driven displacement and interventions to support employment, recognizing that AI's effects extend beyond software capability to jobs and economic policy. By focusing on specific tasks where AI can assist, improve processes, or reduce repetitive effort, businesses can adopt AI in a practical and reviewable manner.

Tracking policy developments that could affect the availability or use of frontier-model features is also recommended for maintaining clarity in AI product choices and public debate.