Top AI Safety Expert Resigns with Enigmatic 'Peril' Warning
Mrinank Sharma, an Oxford alumnus who directed the Claude chatbot developer's Safeguards Research Team, published his departure notice on X Monday, detailing an intensifying personal confrontation 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 to colleagues.
The exit emerges against escalating friction around the San Francisco-headquartered AI laboratory, which finds itself simultaneously accelerating development of increasingly capable systems even as its own leadership cautions those identical technologies could endanger humanity.
The resignation also surfaces following accounts of deepening discord between Anthropic and the Pentagon regarding the military's ambition to utilize AI for autonomous weapons targeting absent the protections the firm has attempted to enforce.
Sharma's departure, landing mere days following Anthropic's launch of Opus 4.6—a more advanced version of its signature Claude platform—suggested internal conflicts regarding safety emphasis.
"Throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions," he wrote. "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."
The scientist's division was created just beyond a year prior with authority to address AI security challenges spanning "model misuse and misalignment," bioterrorism deterrence, and "catastrophe prevention."
Sharma highlighted with satisfaction his contributions building protections against AI-enabled bioweapons and his "final project on understanding how AI assistants could make us less human or distort our humanity." He now plans returning to the UK to "explore a poetry degree" and "become invisible for a period of time."
Anthropic CEO Dario Amodei has persistently cautioned about threats created by the precise technology his enterprise is marketing. In a nearly 20,000-word manifesto last month, he alerted that AI systems of "almost unimaginable power" are "imminent" and will "test who we are as a species."
Amodei flagged "autonomy risks" where AI could "go rogue and overpower humanity," and proposed the technology might facilitate "a global totalitarian dictatorship" via AI-driven surveillance and autonomous weaponry.
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
