2025_CSET对美国人工智能行动计划的建议（英文版）_18页_1mb

US AI Leadership Recommendations

• Increase public sector research funding for AI applications in social and life sciences, with broad sharing of results to foster innovation and discovery.
• Lower barriers for emerging AI companies to stimulate competition and diverse methods in research.
• Develop AI talent through expanded training programs, scholarships, and education initiatives to enhance the workforce and commercial opportunities.

Addressing US-China AI Competition

• Strengthen export controls to prevent illegal technology transfer to China, including scenario planning and enforcement collaboration with allies.
• Enhance intelligence sharing, including open-source intelligence (OSINT) collection and cross-sector information exchange.
• Implement AI event reporting systems to track failures, improve safety, and promote transparency.

Mitigating AI Risks

• Protect the public from AI harms by establishing channels for challenging AI decisions and safeguarding whistleblowers against retaliation.
• Counter AI-enabled biological risks through integrated security measures, threat assessments, and model evaluation.
• Advance AI evaluation science and standards to ensure reliability, reduce risks, and support standardized adoption across sectors.
• Develop comprehensive standards for AI assessment, including cybersecurity and biometric safety, to facilitate fair comparisons and international cooperation.

This summary highlights key recommendations from CSET's formal response, focusing on enhancing US AI leadership, managing competition, and reducing risks.