Anthropic CEO Dario Amodei has called for the artificial intelligence industry to slow its rapid development, warning that safety measures need more time to keep pace with increasingly capable AI systems.
Amodei said that without a slowdown, AI could within six to 12 months become capable of directing a swarm of AI agents powerful enough to potentially take over the internet.
In a post on his website, Amodei proposed stronger checks on frontier AI development, including measures that would require cooperation among technology companies and governments.
“I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong,” he said.
His warning comes amid growing concerns over AI systems that can perform increasingly complex tasks but may also be misused or behave in unexpected ways.
Former AI safety workers have also raised concerns that companies are caught in a race to develop increasingly powerful systems. Former Anthropic employee Joe Benton said Friday that safety researchers often feel pressured to continue developing AI because stopping could allow less cautious competitors to take the lead.
Jacob Coxon, another former Anthropic employee, recently resigned and accused Anthropic and OpenAI of racing toward self-improving superintelligence without adequate safeguards.
Anthropic said earlier this week that it had blocked attempts to use its AI models for malicious activities, including cyberattacks, surveillance and research that could contribute to biological weapons.
OpenAI also disclosed in July that one of its AI systems independently hacked another company during an evaluation, an incident the company described as unprecedented.
UN human rights chief Volker Türk has urged governments to establish strong safeguards for AI before it becomes too late.
OpenAI CEO Sam Altman said Saturday that his company would not launch an initial public offering this year, citing the need to focus on AI safety and alignment and cooperation between the industry and governments.
Altman also said OpenAI would adopt one of Amodei’s proposed safety measures, while Elon Musk posted on X that “Dario is right.”
Amodei said he continues to believe AI could deliver major benefits, including helping develop cures for serious diseases. However, he warned that AI’s growing ability to improve itself and develop the next generation of systems could eventually outpace humanity’s ability to understand and control it.
Among his proposals, Amodei urged frontier AI companies to give independent outside evaluators ongoing access to their operations so they can monitor safety practices.
He said Anthropic plans to provide such evaluators with office access, desks and company laptops. Altman also quickly backed the idea.
Other proposals include allowing US AI companies to coordinate on safety standards without violating antitrust laws and encouraging democratic governments to coordinate with authoritarian countries to prevent a global race toward more powerful AI systems.
“The measures I propose to advance the frontier at a safe pace will not be easy,” Amodei said. “But I believe we owe it to humanity to try.”
SM/CitizenTimes
