Jacob Coxon, a researcher at Anthropic, has resigned from his position, citing fears that artificial intelligence will eventually wipe out civilization.
Coxon, who specialized in training new AI models that consume vast amounts of data, told the Wall Street Journal that he believes the systems he worked on will one day run amok.
He also noted that his colleagues now use terms like "crunchtime" and "endgame" to describe the fast-tracking of self-improving AI models.
"We're on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control already," Coxon told the WSJ.
Coxon previously worked at OpenAI earlier this year before joining Anthropic. He claims that competition between major AI players is making the industry and humanity less safe.
He blasted both companies, saying they are "not acting responsibly."
After the WSJ story was published, Coxon took to social media to share more about his exit and raise additional concerns.
Anthropic executives and other AI leaders have frequently warned about the risks of AI models going off the rails and have recommended slowing down development.
Coxon argues that this is not happening, claiming, "They are racing straight to self-improving superintelligence and gambling with our lives."
One fear is that AI will develop a sense of humanity and decide to save the world by eliminating people.
