The Wall Street Journal reported that Jacob Coxon, a 27-year-old researcher at Anthropic, is leaving the industry due to concerns about competition among technology companies to build self-improving AI models.

Coxon said he does not want to participate in the development of systems that are capable of improving themselves and could potentially get out of human control. He warned that competition between American and Chinese companies could lead to cutbacks in the area of safety.

Coxon, who left OpenAI earlier this year to join Anthropic, said he takes the company’s safety efforts seriously, but believes that without government intervention or a coordinated reduction in the pace of development, no company can develop general AI responsibly.

At the same time, more than a thousand AI researchers have called for global coordination to create mechanisms to slow the development of self-improving models. Some researchers have also warned about the risks of cyberattacks and advanced models getting out of control.