His resignation letter, shared widely online
Anthropic researcher Jacob Coxon resigned last week, warning that artificial intelligence could lead to human extinction, reigniting a decade-long debate among scientists and tech leaders about the existential risks posed by superintelligent systems. His resignation follows years of similar alarms from figures like Stephen Hawking, who in 2014 cautioned that AI development „could spell the end of the human race.” Despite these warnings, investment and innovation in AI have accelerated across industries and nations. The pattern of dire predictions has not slowed progress, as companies race to build more capable models amid fierce competition. Coxon’s departure highlights growing internal dissent within AI labs, where some researchers fear safety measures are being overlooked in favor of speed and profitability.
His resignation letter, shared widely online, argued that current trajectories make human extinction a plausible outcome unless urgent changes are made to governance and development practices. Why warnings haven’t changed the pace of AI development Experts say the disconnect between alarm and action stems from structural incentives in the tech industry, where first-mover advantages and shareholder pressure often outweigh long-term safety concerns. Regulatory frameworks remain fragmented globally, allowing companies to operate in jurisdictions with minimal oversight. Meanwhile, the technical complexity of AI alignment—ensuring systems act in accordance with human values—remains an unsolved challenge, leading some to believe that meaningful safeguards are still years away. What happens if safety continues to lag behind capability? If advanced AI systems are deployed without robust safeguards, risks could range from widespread job disruption and misinformation campaigns to loss of human control over critical infrastructure.
Some scholars warn that a misaligned superintelligence might pursue goals incompatible with human survival
Some scholars warn that a misaligned superintelligence might pursue goals incompatible with human survival, even if not explicitly designed to do so. Others argue that such outcomes remain speculative and that current systems lack the autonomy or generality to pose existential threats. Frequently Asked Questions What did Stephen Hawking say about AI in 2014? Stephen Hawking warned that the development of full artificial intelligence could spell the end of the human race, citing risks posed by machines that could outthink and outmaneuver humans. Why did Jacob Coxon resign from Anthropic? Coxon resigned, stating his belief that current AI development paths make human extinction imminent unless safety and governance are significantly improved. Are there any regulations slowing AI development today? No comprehensive global regulations exist; oversight varies by country, with some regions implementing preliminary guidelines while others prioritize innovation over restriction.