Anthropic Safety Lead Says AI 'Could Kill All Humans' Probability Exceeds 10%, Colleague's Resignation Draws Attention
A senior safety researcher at Anthropic says there is a more than 10% chance AI 'could kill all humans' by the end of the decade. The remark came hours after his colleague Jacob Coxon resigned over concerns that the company and its competitors are neglecting safety while developing uncontrollable 'superhuman systems.'
Anthropic Safety Lead Says AI 'Could Kill All Humans' Probability Exceeds 10%, Colleague's Resignation Draws Attention
A senior safety researcher at Anthropic said there is a more than 10% chance that artificial intelligence "could kill all humans" by the end of this decade. The remark came hours after his colleague Jacob Coxon resigned over concerns that the company and its competitors are neglecting safety while developing uncontrollable "superhuman systems."
Coxon, who has trained AI systems at both Anthropic and OpenAI, announced his departure on X and criticized the company's lax safety measures. According to his statement, he believes AI companies are "gambling with our lives" as they race to develop self-improving AI.
The Anthropic safety lead, who was not named, stressed that the risk must be taken seriously. The episode highlights internal disputes in the AI safety field, especially amid rapid technological iteration, as discussions on regulation and ethics become increasingly urgent. Anthropic did not immediately respond to a request for comment.
Source: https://www.theverge.com/ai-artificial-intelligence/991927/anthropic-ai-kill-all-humans
Provenance & status
- Byline
- OceanAlt Editorial
- First published
- 2026-09-10
- Last updated
- 2026-09-10
- Content type
- Newsflash
- Source material
- View original ↗
Related reading
Binance Sets Aside $6.34B to Cover Hack Losses; Attack Chain Exposes Account Abstraction Authorization Blind Spot
Beyond the Binance Security Incident: Why AI Payments Need a New Line of Defense
Harbor Enters AI Agent Insurance With $10M Seed Round, Leaving Liability Attribution and Claims Pricing as the First Unsolved Problems
Paste a payee address before you pay and see whether it's on a sanctions list, through a mixer, or tagged for fraud.
This judgement can sit inside your own product
One line of code; it touches neither your CSS nor your JS. The same pre-settlement judgement can appear in your articles, on your wallet's confirmation screen, or as an endpoint your agent calls before it pays.