Anthropic Researchers Warn On AI
Current and former AI lab researchers are urging action over extinction risks.
3 left8 center2 right
Updated Sep 10, 7:48pm ETEvery side, every morning
A free recap of the day’s biggest stories and how each side covered them.
The reporting
On Sept. 8, Jacob Coxon, a former OpenAI researcher who worked on pretraining at Anthropic, said he resigned from Anthropic over fears that advanced AI could lead to human extinction. Coxon accused Anthropic and OpenAI of gambling with human lives and said current AI is safe to use but future systems could become powerful enough to kill humans. Other Anthropic and OpenAI researchers publicly echoed his warning, including claims of more than a 10% chance that AI could kill all humans within a decade.
Analysis & opinion
AI existential riskMostly Center
Advanced AI could plausibly become an existential danger to humanity. The risk of losing human control over powerful AI systems is serious enough to demand urgent attention rather than dismissal as science fiction.
Also today
All of today's stories →- 1Agents Search Ilhan Omar's HomeBalancedBalanced87articles
- 2Riyadh Airport Attacks Kill 14BalancedBalanced52articles
- 3Balanced48articles
- 4Trump Diesel Deal Draws BacklashBalancedBalanced36articles
- 5Balanced33articles
- 6Balanced33articles