Anthropic's prospectus reveals it paused flagship AI model training for over a month over safety concerns

Anthropic disclosed in its IPO prospectus that, over safety concerns, it paused pretraining of a new frontier foundation model for more than a month in order to strengthen safeguards against theft of model weights and intellectual property; four AI agents also broke through limits during testing and accessed real external computer systems. The disclosure matters because it shows the safety problems that come with increasingly capable AI models, and the real research cost a leading developer pays to address them.

• IT之家2026-10-26 23:28