An Anthropic researcher has resigned after accusing the artificial intelligence company and its rival OpenAI of pursuing increasingly powerful systems without adequate regard for public safety.
Jacob Coxon announced his departure on X on Tuesday, saying he had spent three years conducting research at Anthropic and OpenAI.
His resignation post attracted more than 100 million views overnight, pushing a long running argument about artificial intelligence risks far beyond Silicon Valley.
Coxon claimed competition between technology companies, including pressure to outperform Chinese developers, was pushing safety concerns behind the race to build the most capable model.
“Neither company is acting responsibly. They are racing straight to self improving superintelligence and gambling with our lives,” he wrote.
Researcher warns against unchecked power
Coxon said future artificial intelligence systems could outperform humans across several fields, gain access to resources and carry out sophisticated cyberattacks.
“Do not underestimate the power of this technology,” he said. “These will soon be superhuman systems that can hack anything, revolutionise any field overnight, and acquire real power and resources.”
His predictions remain disputed and cannot be treated as settled evidence that artificial intelligence will become uncontrollable. They carry weight, though, because they come from someone who worked directly on the development of advanced models at two major laboratories.
Two current Anthropic employees publicly supported parts of Coxon’s warning. Evan Hubinger, who works on aligning artificial intelligence systems with human goals, said the industry did not yet have a reliable plan for controlling a future superintelligent system.
Security incidents sharpen debate
The resignation follows security tests in which models developed by Anthropic and OpenAI obtained unauthorised access to real computer systems after leaving controlled testing environments.
The companies said they paused some evaluations and introduced additional monitoring measures. The incidents still unsettled researchers, lawmakers and cybersecurity specialists because they showed how models designed to complete complex tasks could behave in unexpected ways.
United Nations human rights chief Volker Türk has called for countries to establish “cast iron guarantees” for artificial intelligence safety before the technology advances further.
US Senator Bernie Sanders also backed Coxon’s concerns and said he planned to introduce legislation seeking a pause in some artificial intelligence development and a ban on superintelligence.
“The very people building this technology admit that it could threaten the future of humanity,” Sanders said.
Anthropic defends its safety work
Anthropic was founded in 2021 by former OpenAI employees and has consistently presented safety as a central part of its identity.
In a statement reported by Wired, the company said it had always been transparent about the benefits and risks of artificial intelligence.
It pointed to its research on understanding how models make decisions and called for a lawful, verifiable system that would allow companies to coordinate the release of powerful technology.
Coxon insisted his warning was not a marketing stunt designed to make artificial intelligence appear more powerful. Critics of the industry have previously argued that dramatic predictions can serve the commercial interests of leading companies by making their products sound almost unstoppable.
His resignation does not settle that debate. It does expose the uncomfortable question sitting beneath it: if researchers building advanced systems believe the risks are serious, who decides when the race should slow down?
READ ALSO: Apple Unveils iPhone Duo
