Anthropic Filing Says It May Soon Hit Its Own AI Acceleration Threshold

Anthropic says it may soon reach a threshold in its own safety policy for AI that speeds up AI research, according to the amended IPO registration statement it filed with the SEC on 2 November. "Our internal estimates and our expectations regarding future models, rather than our current models, indicate that we may imminently meet our Responsible Scaling Policy's definition of 'dramatically accelerated progress attributable to AI,'" the filing says.

The company adds that it has "not designated any of our current models as exceeding this threshold" and continues to assess the underlying data. It says it is evaluating what meeting the AI research and development capability threshold would mean and will publish further reports as its policy requires.

The passage is new. The original S-1 filed on 26 October warned that Anthropic "may not identify a significant acceleration in our rate of development until after it has occurred." The amendment now says the company has "in the past not identified" such acceleration until after the fact.

Both versions say its use of AI to speed up its own research may push model capabilities ahead of its safety, alignment and security measures, and that precautions have included delaying releases, as with Mythos through Project Glasswing. The same filing reported third quarter revenue of $17.0 billion.