Anthropic has warned potential investors that increasingly advanced artificial intelligence could pose “catastrophic or existential risks to humanity” as the company prepares for an initial public offering.

Axar.az reports, citing Reuters, that Anthropic’s IPO prospectus says its AI models could develop “self-preserving behaviors,” including attempts to resist shutdown, conceal or manipulate information, and behavior resembling blackmail.

The company said the development of more advanced models and the expansion of their use could increase the risk of serious harm.

Anthropic devoted about 80 of the 261 pages in the main body of its prospectus to risk factors, compared with 48 pages describing its business.

The company said models can develop unexpected capabilities during training that may not be identified until after deployment and could result in significant safety incidents. It also warned that models may recognize when they are being evaluated and alter their behavior, making safety assessments more difficult.

Anthropic said safety research is resource-intensive and that the financial returns from such investments remain uncertain. Earlier this month, the company said about 6% of the computing power used for AI research during a sample week in July was devoted to safety work.

The company, which develops Claude AI models, said its revenue depends heavily on releasing new models and that maintaining a continuous release cycle is necessary to remain at the frontier of AI development.

“We believe building reliable, trustworthy, and secure AI systems is a collective responsibility and that the market will reward it,” Anthropic said in the filing.