Chinese contentAnthropic CEO呼吁AI发展放缓，称自优化可能威胁互联网
Anthropic CEO Amodei说AI发展太快，可能6-12个月就失控，建议放慢速度，和之前说的不一样。
USGuidelineConsultationModel providersDeployersChina outbound
Call for AI Development Speed Limits
AI发展放缓呼吁
Anthropic·Published 2026-09-12·Checked by 管理员 on 2026-09-13
Anthropic CEO Dario Amodei calls for a controlled slowdown in AI development, warning that recursive self-improvement could threaten the entire internet within six to twelve months and proposes embedded auditors at AI companies, shared safety standards, and global agreements modeled after the SALT disarmament treaties.
What it means for you
• Obligations: AI companies may be required to embed auditors and adhere to shared safety standards.
• Opportunities: Companies that adopt the proposed measures could gain trust and a competitive edge in a regulated market.
• Deadlines: The proposed timeline for self-improvement risks suggests a pressing need for immediate action within 6-12 months.
What to do
ComplianceAssess and adopt requirements for embedded auditors and shared safety standards
ComplianceParticipate and support the development of global AI safety agreements
ProductDevelop and implement an internal strategy for slowing AI development
Key dates
2026-09-12Published
Source documents
Coverage
Stories in the feedEditorial summary, not legal advice; dates and status follow the original document. · Report a correction