Ex-Anthropic Researcher Warns AI Labs Have Years to Make Systems Safe

Jacob Coxon, a researcher who recently left AI company Anthropic, has spoken to WIRED about his concerns over the direction of AI development. He described internal efforts at Anthropic as resembling a 'mini Manhattan Project,' drawing parallels to the high-stakes race to develop nuclear weapons. Coxon raised serious doubts about the effectiveness of current AI alignment approaches, which aim to ensure AI systems behave safely and as intended. He warned that AI laboratories have only a limited window of a few years to address these safety challenges before they become unmanageable. His departure and public comments reflect growing tensions within the AI industry over how urgently safety concerns are being treated.