Story perspectives

Anthropic Warns IPO Prospectus of Existential AI Risks

By Drooid · Sep 29, 2026, 4:05 PM UTC · How we work

Summary

Story summary

Anthropic warned its AI models could cause catastrophic or existential risks in its IPO prospectus.
The prospectus devotes nearly 80 of 261 pages to risk factors, twice the length of the business section.
Safety researcher Evan Hubinger estimated AI has over a 10% chance of killing humans within a decade.
President Donald Trump rejected the slowdown call, urging the United States to win the AI race.

History

Historical Context

In 2015, OpenAI was launched as a nonprofit with a charter pledging $1 billion to develop artificial general intelligence safely, echoing a January 2015 open letter signed by Elon Musk and Stephen Hawking warning of existential risk. The organization’s safety focus spurred the first coordinated AI-risk discourse, a trend that now drives Trump’s AI-race push and the Senate’s review of the 2020 National AI Initiative Act.

1 of 2

Story summary

Anthropic warned its AI models could cause catastrophic or existential risks in its IPO prospectus.
The prospectus devotes nearly 80 of 261 pages to risk factors, twice the length of the business section.
Safety researcher Evan Hubinger estimated AI has over a 10% chance of killing humans within a decade.
President Donald Trump rejected the slowdown call, urging the United States to win the AI race.

1 / 2