Why an Anthropic AI Researcher Resigned Over Safety Concerns
An Anthropic AI researcher named Jacob Coxon has resigned from the company, delivering a stark warning that artificial intelligence could spiral out of control by the end of next year and potentially destroy humanity. Coxon, a 27-year-old British researcher specializing in large-scale data training for new AI models, shared his concerns on X (formerly Twitter) on September 8, 2026. He pointed out that the intense AI development race is sidelining critical safety protocols.
Internal Executives Confirm the Threat of Uncontrollable AI
According to reports from The Wall Street Journal, Coxon stated that many of the most aggressive and dangerous AI scenarios are actively manifesting. Anthropic’s Head of Alignment, Evan Hubinger, publicly agreed with Coxon, noting that executives genuinely believe advanced AI systems could pose an existential threat to all of humanity, estimating the probability at over 10 percent within the decade. Samuel Marks, Head of Expansion Oversight, also echoed that developers acknowledge the risk of catastrophic outcomes.
The Call for Government Intervention and Global Regulation
Coxon previously worked at OpenAI before moving to Anthropic earlier this year in search of a stronger safety culture. However, he concluded that individual company efforts are insufficient and that both OpenAI and Anthropic are gambling with human lives in their race toward self-improving artificial general intelligence (AGI). Coxon and roughly 1,000 other researchersβincluding Anthropic CEO Dario Amodeiβhave previously signed statements urging global governments to establish international regulatory frameworks to slow down uncontrolled AI development.
