Evan Hubinger, who works at Anthropic on aligning artificial intelligence with human values, said he believes there is a more than 10 percent chance that AI will kill all humans within 10 years.
Hubinger responded to a post by Jacob Coxon, who resigned from Anthropic, claiming that AI developers believe the technology could kill everyone by the end of the next 10 years. Hubinger said he agreed with Coxon’s assessment and that they genuinely believe this possibility exists.
No plan exists to align superintelligence
Hubinger said he believed Anthropic was doing everything it could, but acknowledged that the company had no clear plan for solving the superintelligence alignment problem. He also said there was no clear roadmap for progress in this area.
Coxon announces his departure from Anthropic
Jacob Coxon announced his resignation from Anthropic, saying he had worked as a researcher at OpenAI and Anthropic over the past three years. Coxon argued that neither company had acted responsibly enough.
Coxon claimed that companies were engaged in an uncontrolled race to achieve self-improving superintelligence. He said these systems could quickly become superhuman entities capable of hacking everything, transforming various fields overnight and gaining access to real power and resources.
Coxon also claimed that workers in the artificial intelligence sector seriously believe these systems could kill humanity within 10 years. He argued that executives use more moderate language in public but experience deep fear behind closed doors.