Anthropic 有信心“在未来几天内”重新开放 Mythos 和 Fable 5 的访问权限
Anthropic 国际业务董事总经理 Chris Ciauri 周三在首尔的新闻发布会上表示，公司非常有信心在未来几天内重新开放其前沿 AI 模型 Mythos 和 Fable 5 的访问权限。此前，白宫指示阻止外国用户使用这些模型。
背景速读
- Anthropic is the American AI company behind the Claude series of large language models, founded by former OpenAI employees. It positions itself as a "safety-first" AI lab.
- "Mythos" and "Fable 5" are widely believed to be internal codenames for Anthropic's most advanced (frontier) AI models — likely the next-generation Claude versions that have not yet been publicly released or were briefly restricted.
- The White House recently directed Anthropic to block access to these models for foreign nationals, likely under national security concerns about advanced AI capabilities being used by adversaries. This is part of a broader U.S. government trend of restricting cutting-edge AI exports.
- Chris Ciauri is Anthropic's Managing Director of International, responsible for the company's global go-to-market strategy and government relations outside the U.S.
- The press conference in Seoul suggests South Korea is a key market or partner for Anthropic, and that the company is actively working with allied governments to resolve access restrictions.