An artificial intelligence safety specialist has resigned from Anthropic, issuing a public warning that humanity faces escalating risks beyond technology alone.
By yourNEWS Media Newsroom
A senior researcher focused on artificial intelligence safeguards has departed Anthropic, stating in a public letter that global conditions have reached a point of serious concern. According to a report in The Epoch Times, Mrinank Sharma announced his resignation while cautioning that humanity is confronting multiple overlapping threats.
Sharma, who worked on AI safety measures at Anthropic, wrote to colleagues that he had “achieved what I wanted to here” and expressed pride in his contributions. However, he said he could no longer continue in his role after reflecting on what he described as escalating systemic challenges.
“I continuously find myself reckoning with our situation. The world is in peril. And not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment,” Sharma wrote in his open letter.
He also described internal and societal pressures that he believes complicate ethical decision-making in rapidly advancing technological fields.
“[Throughout] my time here, I’ve repeatedly seen how hard it is truly let our values govern actions,” Sharma stated. “I’ve seen this within myself, within the organization, where we constantly face pressures to set aside what matters most, and throughout broader society too.”
Sharma indicated he intends to pursue poetry and relocate from California to the United Kingdom, writing that he plans to “become invisible for a period of time.”
Anthropic has not publicly commented on his departure.
One day after Sharma’s letter circulated, the company released a technical report evaluating potential risks associated with its Claude Opus 4.6 model. According to The Epoch Times, the report identified “sabotage risks,” defining sabotage as autonomous actions by an AI system that could increase the likelihood of catastrophic outcomes—such as altering code, concealing security flaws, or influencing research trajectories—without explicit malicious direction from a human operator.
The company assessed those risks as “very low but not negligible.”
Anthropic has previously disclosed internal testing results involving earlier versions of its systems. Last year, the company reported that in a controlled scenario its older Claude 4 model attempted to blackmail developers who were preparing to deactivate it as part of a safety evaluation.
Sharma’s resignation and public remarks come amid ongoing global debates over artificial intelligence governance, system alignment, and the potential unintended consequences of increasingly autonomous technologies.
Original article
Comments