A senior researcher at artificial intelligence company Anthropic has warned that he believes there is more than a 10% chance AI could kill all humans within the next decade.
Evan Hubinger, who leads Anthropic’s alignment science efforts, made the assessment on social media while responding to fellow researcher Jacob Coxon, who announced his resignation from the company over concerns about the safety of advanced AI development.
Coxon accused leading AI companies of racing toward self-improving superintelligence without adequate safeguards, warning that increasingly powerful systems could eventually acquire capabilities beyond human control.
Hubinger agreed with the concerns, saying that Anthropic researchers genuinely believe AI could pose an existential threat to humanity. He personally estimated the likelihood of AI killing all humans at more than 10% over the next decade.
However, Hubinger stressed that the risk posed by AI models currently in use is low. His concern is primarily focused on future systems that could potentially improve themselves and develop capabilities far beyond today’s models.
China Warns of ‘Security Backdoor’ in Anthropic’s Claude Code AI Tool
No clear solution for superintelligence
Hubinger also acknowledged that Anthropic does not yet have a definitive solution to the AI alignment problem for superintelligent systems.
AI alignment refers to efforts to ensure that advanced AI systems behave in accordance with human intentions, values and safety requirements.
The researcher said Anthropic was trying to address the issue but was not clearly on track to develop a solution for superintelligence.
Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing.
— Jacob Coxon (@hilbertspaess) September 9, 2026
The warning came as concerns over the rapid development of advanced AI systems grow within the industry. Coxon’s resignation has added to a broader debate over whether AI companies are moving faster than their ability to understand, monitor and control increasingly capable systems.
Experts have warned that one of the major future risks could emerge if AI systems gain the ability to improve their own capabilities, potentially creating a cycle of rapid development that becomes increasingly difficult for humans to control.
Hubinger’s comments have renewed debate over how AI companies should balance rapid technological progress with safety research and safeguards designed to prevent catastrophic outcomes.
