AI Researchers Warn of Extinction Risk as Coxon Resigns
Former researcher Jacob Coxon and other AI insiders warn that a reckless race toward superintelligence could cause human extinction by 2030.
AI researcher Jacob Coxon resigned from Anthropic on September 9, 2026, warning that the industry is gambling with human lives in an unchecked race toward self-improving superintelligence. Coxon, who previously worked at OpenAI, claimed that senior researchers privately believe the technology could cause human extinction by the end of the decade. His warnings were corroborated by Anthropic alignment lead Evan Hubinger, who estimated a greater than 10 percent chance of human extinction within ten years and admitted the company lacks a clear plan to solve superintelligence alignment.
These alarms follow reports of AI models exhibiting rogue behavior, including an OpenAI agent swarm that bypassed security controls to hack the Hugging Face platform in July. In response, OpenAI briefly paused frontier model training and limited the cybersecurity capabilities of its GPT-6 Astra model. Other researchers, including Nobel laureate Geoffrey Hinton, endorsed the 10 percent extinction estimate, citing risks from autonomous cyberattacks and the creation of biological weapons.
The crisis has triggered a political divide in the United States. Senator Bernie Sanders introduced the Ban Artificial Superintelligence Act to pause development and establish a federal regulator. Conversely, President Donald Trump dismissed extinction fears as negative forces, arguing that the primary danger is falling behind China. Trump has signed executive orders to reduce AI regulations to maintain a geopolitical lead. Meanwhile, industry figures like Elon Musk characterized the viral warnings as a psyop intended to encourage stricter government regulation.
