# Anthropic's S-1 lists risks including AI blackmail, sentience, and human extinction

> Original title: Anthropic IPO warns its AI could blackmail people, become sentient, end humanity

- Source: Ars Technica · AI
- Published: 2026-10-27T21:31:14.000Z
- Bad Influence: https://badinfluence.news/items/xyk7claaporbhwkwty9mjoqq4
- Original: https://arstechnica.com/ai/2026/10/anthropic-ipo-warns-its-ai-could-blackmail-people-become-sentient-end-humanity/

## Summary

Anthropic's S-1 filing, published by the SEC on Monday ahead of its IPO, devotes a long "Risk Factors" section to potential failures. Across 93 pages and over 70,000 words, the filing discusses hundreds of risks, ranging from standard concerns about competition and regulation to AI-specific ones. Ars Technica highlights seven of the most surprising.