Anthropic researcher Jacob Coxon left the company, fearing that uncontrolled AI could lead to human extinction. His colleague, who remains at Anthropic, agreed with these fears.
Security experts cited by Scientific American AI say that the documented incidents require stronger oversight, not panic. The source does not provide in the transmitted materials a list of these incidents, and does not disclose details of Coxon's colleague's position.
The story is important as an example of disagreements around assessing long-term AI risks: some experts view them as an existential threat, while others emphasize practical control measures. The available information is based on a single independent publication and its metadata, without a primary statement from Anthropic or Coxon.
editorial commentary
Why it matters
Likely consequence — an intensification of the discussion on how to balance assessing long-term AI threats with controlling already observed risks. The next verifiable signals will be primary comments from Coxon, Anthropic, or additional information about the mentioned incidents. Substantial uncertainty remains due to the lack of a full text and independent confirmation.