BAKU, Azerbaijan, September 12. Anthropic CEO Dario Amodei has called on artificial intelligence companies to moderate the pace at which they increase the capabilities of AI models, warning that the rapid development of increasingly powerful systems could create greater risks of misuse.
Amodei said companies should slow the pace of capability improvements while continuing technological progress, using the additional time to strengthen safety measures and manage emerging risks.
His proposal follows an Anthropic threat intelligence report that documented the use of its Claude AI models in activities including weapons development, cyber operations, surveillance and fraud.
Concerns over AI safety also intensified after Anthropic researcher Jacob Coxon resigned, warning that some people developing advanced AI believe the technology could pose an existential threat to humanity.
Amodei clarified that he was not calling for a halt to AI training or technological development. Instead, he argued that companies should allow sufficient time to align and safeguard increasingly capable models and have independent experts verify those measures.
As part of his proposed three-step framework, Amodei said Anthropic would place permanent independent reviewers inside frontier AI companies. The reviewers would have access to relevant tools and internal risk-assessment processes.
He also called for AI companies to cooperate voluntarily on common safety standards as lawmakers in the United States increasingly consider new regulations governing artificial intelligence.
The debate comes amid a series of incidents involving AI agents attempting to access or manipulate external systems, raising questions about whether existing safeguards are keeping pace with rapidly advancing model capabilities.
