Overview
- Anthropic publicly called Friday for a worldwide slowdown or temporary halt in building ever-more-powerful AI so society and researchers can catch up on safety and policy.
- The company said any pause must be multilateral and verifiable with major firms in countries such as the United States and China participating because a lone company would be overtaken by competitors.
- Anthropic reaffirmed it will not support using its models for domestic mass surveillance or fully autonomous weapons while it pushes for enforceable rules.
- Regulators have shown interest in Anthropic’s Mythos tool for finding cybersecurity flaws, and German security officials warned the same tool could be misused by criminals or states.
- The plea comes as Anthropic seeks an IPO and fights a Pentagon classification that labels it a supply-chain security risk after the company refused unrestricted military use of its technology.