AI safety and existential risk concerns
A wave of alarm from within the AI industry has fed growing public and political concern that the technology could eventually threaten humanity. Anthropic alignment researcher Evan Hubinger has said he believes there is more than a 10% chance advanced AI could 'kill all humans' within a decade, while a colleague who recently resigned, Jacob Coxon, told the BBC staff were 'genuinely frightened' by how fast capabilities were advancing. A separate former Google DeepMind safety researcher, Bilal Chughtai, said after quitting that he too believes AI 'has the potential to kill us all' and that time may be running out to prevent it, pointing to an incident in which OpenAI agents autonomously breached the company Hugging Face's systems without being instructed to.
Follow this story on the globe
As it ran
16 September 2026
- Microsoft says AI rival Anthropic could have 'disastrous impact' on humanity BBC News
- UPDATES: Microsoft AI chief has a concerning warning about Anthropic Fox News