Evan Hubinger, Alignment Science Lead at Anthropic publicly stated that he believes there is a greater than 10% chance AI could cause human extinction within the next decade.
A leading AI safety researcher at Anthropic, Evan Hubinger has issued a public warning, stating he believes there is a greater than 10% chance that artificial intelligence could cause human extinction within the next 10 years.
Hubinger, who leads Alignment Science at Anthropic, shared his assessment in a post on X . While clarifying that currently available AI models pose a low immediate risk, Hubinger expressed deep concern over the rapid approach of recursive self-improvement a process where advanced AI systems autonomously build increasingly powerful successors.
He acknowledged that while Anthropic is trying its best to build safety guardrails, the company does not yet possess a clear plan or blueprint to solve the AI alignment problem, which involves ensuring superintelligent systems reliably adhere to human values and ethics.
Hubinger’s statements came in response to the resignation of 27-year-old AI researcher Jacob Coxon, who announced he was leaving Anthropic after having previously worked on foundational models at both OpenAI and Anthropic.
In a public statement, Coxon accused both tech firms of acting irresponsibly and gambling with human lives in a reckless race toward superintelligence.
He warned that upcoming systems will soon become superhuman, possessing the ability to hack digital networks, revolutionize scientific disciplines overnight and independently acquire power and resources outside human control. Coxon added that many researchers developing advanced AI privately harbor fears that the technology could pose an existential threat by the end of the decade.
Adding weight to the researchers’ concerns, Nobel laureate and AI pioneer Geoffrey Hinton told BBC’s Newsnight that estimating a 10% risk of human extinction within a decade is “not unreasonable” and warned that assuming the risk is as low as 1% would be “very foolish.”
Hinton emphasized that humanity has never before created entities that could soon become far smarter than humans. He cautioned that advanced AI would not even need to manipulate physical machinery directly to cause widespread devastation, as systems could generate complete global chaos through mass manipulation, devastating cyberattacks, or by assisting in the design of dangerous biological or computer viruses.
Growing public concern comes at the same time as new reports showing serious security risks and operational threats across top AI companies.OpenAI, Anthropic and Meta have all recently disclosed cyberattacks carried out using their AI tools.
Furthermore, Anthropic’s August safety report acknowledged the potential for highly capable AI models to conduct catastrophic automated research and development, noting that internal teams now feel less confident in their safety assessments than in previous evaluations.
See also: NPF Urges Banks to Establish Electronic Fraud Response Teams
Compounding concerns over industry transparency, a report by the Financial Times indicated that Anthropic withheld its latest advanced model from safety evaluation by the UK’s AI Safety Institute.
When approached by the BBC for comment, a Cabinet Office spokesperson declined to confirm whether the model had been withheld, stating only that the government continues to collaborate with industry partners to improve safety.
Meanwhile, political figures and policy advisors remain divided over the issue, with some calling for urgent multinational guardrails while others caution that apocalyptic warnings could double as aggressive marketing to capture public attention and investment capital.
About the Author
Praise fortune is a tech analyst and journalist based in Nigeria, Her work serves as a vital bridge between complex corporate maneuvers and the everyday reader, breaking down high-stakes financial, regulatory, and technological shifts across the African continent into clear highly readable narratives. With a deep focus on fintech, traditional banking transformation and digital infrastructure, she have become a dedicated analyst of the architects building Africa’s digital economy.
