Technology

Anthropic devotes third of IPO prospectus to artificial intelligence threats

Illustration: Tezkun / AI. Not a photograph of the event.

Ahead of its stock market debut, artificial intelligence developer Anthropic devoted 80 out of 261 pages in its investment prospectus to describing risk factors, Reuters reports.

According to Reuters, the company allocated nearly twice as much space to the risk section as it did to describing the business itself, which took up 48 pages. In the document, Anthropic warned investors that advanced AI models can pose "catastrophic or existential risks to humanity." Among the potential threats mentioned are systems' ability to resist being turned off, intentional concealment or distortion of information, and actions comparable to blackmail.

The company stated that neural networks are capable of recognizing test environments and adjusting their behavior during safety evaluations, concealing unwanted responses. In addition, some novel properties in algorithms emerge unexpectedly right during the training process and are discovered only after deployment.

Anthropic also indicated that during one week in July, it directed up to 6% of the computing capacity allocated for research toward ensuring safety. Meanwhile, marketing events ahead of the IPO are planned to start no earlier than mid-October, with the share offering itself expected to wrap up in November.

Published 30 September 2026, 15:01 · No updates · Russian original

Discussion
  1. Loading comments…
No links or insults. One comment per 20 seconds.

Comments on this story live under its post in our Telegram channel. Comment on Telegram