Top AI Labs Call For Slowdown As Global Leaders Split Over Response

Overview

• On Saturday, September 12, Anthropic CEO Dario Amodei published a long essay urging a slowdown in developing the most capable AI models and proposing independent third‑party evaluators to verify safety.
• Senior executives at OpenAI, xAI and Google DeepMind publicly backed Amodei’s caution and said they will adopt or discuss independent evaluators and other joint safety steps.
• The international response has fractured: the United Nations demanded urgent multilateral regulation, the Chinese state press framed the appeal as geopolitical containment, and President Donald Trump dismissed calls for restraint as "forces negative."
• Reports of model misuse and a coordinated intrusion that targeted developer infrastructure, plus a researcher’s high‑profile resignation warning of existential risk, have driven companies and investors to treat safety as an urgent priority and helped trigger declines in tech and chip stocks.
• Practical work is under way—companies are holding talks and Microsoft proposed a code of conduct—but antitrust doubts, verification challenges and short legislative calendars mean binding global rules are unlikely in the near term.