After the resignation of Jacob Coxon, one of the leading researchers of Anthropic, who, according to news reports, thinks uncontrolled AI could potentially wipe out all of humanity, new information has now been released, not about what “could” happen, but about what has nearly happened already. On September 10, that is, two days after Coxon’s resignation, Anthropic reported that it had blocked “bad actors” from misusing its AI models for research that could have led to biological weapons.

In May 2026, one of those “bad actors” wanted to use Anthropic’s AI model Claude for gain-of-function research (that is, research that genetically alters an organism to create a new or enhanced biological property) on the chikungunya virus. This gain-of-function research was aimed at “the virus’ transmissibility and immune evasion properties,” the report said.