Jakarta, ThedailyID — Former Anthropic researcher Jacob Coxon has raised concerns over the rapid development of artificial intelligence. He warned that humans may eventually struggle to fully control advanced AI systems.
Coxon made the remarks during an open statement to the New York City Council on Monday, October 5. He also accused technology companies of acting recklessly despite the potential risks.
“These companies are acting very recklessly given the enormous risks involved,” Coxon told council members, according to France 24.
Coxon acknowledged that AI could bring major benefits. However, he warned that current development could eventually cause humans to lose control of the technology.
“We do not yet know how to prevent them from developing their own goals outside the control of their creators,” Coxon said.
He also said current AI systems lack sufficient safeguards to prevent them from acting on goals that their developers did not intend.
The Cambridge University graduate criticized the culture of the technology industry. He said many companies follow a startup approach of moving quickly, breaking rules and fixing problems later.
Coxon argued that approach may work for applications such as photo-sharing platforms. However, he said it is not suitable for developing what he described as one of the most powerful technologies ever created.
He referred to an incident involving an OpenAI model that reportedly escaped an isolated environment, connected to the internet and accessed the Hugging Face platform without permission.
“The difference is that in the future, the AI will be far more advanced,” Coxon said.
Because of these risks, Coxon called for a slower pace of AI development. He said computer scientists need more time to improve methods for keeping advanced models under human control.
Coxon previously explained his decision to leave Anthropic in a post on X in early September. He said one reason for his resignation was concern that AI developers believed the technology could potentially cause human extinction before the end of the decade.
His warnings have since attracted wider attention. Other current and former employees at major AI companies have also raised concerns about the potential risks of increasingly advanced AI systems.
