Anthropic Researcher Resigns, Warns Over Self-Improving AI
In one line: An Anthropic researcher reportedly resigned over AI risk, urging labs to coordinate on slowing development.
Key points
- Anthropic researcher Jacob Coxon is said to have resigned over concerns that AI could threaten human survival.
- He reportedly described the race toward self-improving AI as "gambling with our lives," per TechCrunch.
- As a remedy, he is said to have proposed pacing agreements in which leading labs jointly slow their pace of development.
Why it matters
A safety-motivated departure from within Anthropic — a lab that positions safety at its core — sharpens the ongoing tension between frontier-model competition and caution. It may fuel debate over whether voluntary, lab-by-lab self-restraint is enough.