AI safety researcher quits with a cryptic warning (www.rt.com)

Mrinank Sharma, a prominent AI safety researcher and leader of the Safeguards Research Team at Anthropic, has resigned, issuing a stark warning about the "interconnected crises" facing humanity today. In his resignation letter, posted on social media, Sharma expressed deep concern over various global threats, not solely from AI, emphasizing that "the world is in peril." His decision to step back from the limelight signals growing tensions within Anthropic as the company grapples with balancing rapid AI development and the ethical implications posed by these technologies. This resignation comes shortly after Anthropic released a new iteration of its Claude chatbot and highlights internal friction regarding safety priorities following reports of discord with the Pentagon over AI for military applications. Sharma's work focused on critical issues such as AI misuse and bioweapon prevention, and his departure underscores the ethical dilemmas faced by AI researchers in an industry focused on innovation yet increasingly aware of its potential dangers. His call for a more cautious and values-driven approach resonates amid CEO Dario Amodei's warnings about the existential risks posed by powerful AI systems.