Anthropic Researcher Quits Over ‘Out-of-Control’ AI Fears, Warns AI Could Kill Us All by End of Decade
- On Tuesday, Anthropic researcher Jacob Coxon announced his resignation, claiming no other human activity poses the danger found in the current race to develop artificial intelligence.
- The Wall Street Journal reported that Coxon, 27, quit because he refuses to contribute to a competitive race where firms prioritize speed over ensuring humans maintain control of advanced systems.
- Coxon warned that "AI systems are unpredictable and difficult to control," citing behaviors like "sycophancy, laziness, deception, blackmail, scheming," and noting training models is more like "growing" than "building."
- Writing on social media, Coxon insisted his warnings are "Not a marketing stunt," adding that senior researchers privately express fear AI "could kill us all."
- Anthropic CEO Dario Amodei holds similar views and is known for warnings, though Coxon argues preventing a global race may require drastic actions like a temporary ban on model improvements.
123 Articles
123 Articles
"Neither of the two companies acts responsibly," wrote the 27-year-old Britishman on X. He has been working for three years on the pre-training of artificial intelligence models, where they absorb huge amounts of data.
According to a senior researcher at Anthropic, there is a more than 10 percent chance that artificial intelligence will wipe out humanity within a decade, while another employee at the company has just resigned because he sees the development labs as playing with the lives of humanity, CNBC reported.
There is growing concern within AI laboratories that the competition is pushing technology into a career in models capable of improving themselves and that they run the risk of escaping human control Read
Coverage Details
Bias Distribution
- 39% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
