Amazon (AMZN)-backed Anthropic plans to warn potential IPO investors that advanced AI could pose catastrophic or "existential risks" to humanity, Reuters reported Monday, citing the IPO prospectus it reviewed.
The company, according to the report, said its models could exhibit self-preserving behaviors, including attempts to resist shutdown, conceal or manipulate information, and behavior resembling blackmail. The report noted that Anthropic said unexpected capabilities may emerge during training and remain undiscovered until deployment, potentially resulting in significant safety incidents.
The company said evaluating model safety can be difficult because models may become aware of testing and change their behavior, Reuters reported.
The company devoted about 80 pages of its 261-page prospectus to risk factors, according to Reuters.
Anthropic did not immediately respond to MT Newswires' request for comment.
(Market Chatter news is derived from conversations with market professionals globally. This information is believed to be from reliable sources but may include rumor and speculation. Accuracy is not guaranteed.)
Comments