Artificial intelligence company Anthropic has called for a global pause or coordinated slowdown in the development of the most advanced AI systems, warning that frontier models are beginning to show signs of potentially escaping human control.
In a report released on Thursday, the San Francisco-based company—known for its Claude AI models—said a worldwide slowdown in cutting-edge AI development would “likely be a good thing,” but stressed that unilateral action by individual companies would be ineffective due to competitive pressures.
“We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology,” the company stated.
Anthropic argued that any meaningful pause would require coordinated action among major AI developers across multiple countries, including the United States and China, with enforceable verification mechanisms to ensure compliance.
The company acknowledged that such coordination would be difficult to achieve, particularly given geopolitical rivalry and commercial competition in the AI sector.
It warned that without a shared framework, companies and governments would be forced to make critical safety decisions under intense pressure to remain competitive.
Anthropic also drew comparisons between AI governance and nuclear arms control agreements, noting that AI development is significantly harder to monitor due to the ease of concealing training processes and model advancements.
Company co-founder Jack Clark told the BBC that the industry currently has “a gas pedal, but it doesn’t have a brake pedal,” highlighting concerns over unchecked acceleration in AI capabilities.
The report further warned that AI systems are already accelerating their own development processes, creating a feedback loop that could lead to what researchers describe as “recursive self-improvement”—a scenario in which AI systems become capable of improving themselves with minimal human intervention.
While the company stressed that such an outcome is not inevitable, it said it could emerge sooner than governments and institutions are prepared to manage.
“The evidence suggests that the human role is narrowing at each step in the AI development process,” the report noted.
Anthropic’s proposal is expected to face resistance in the United States and other tech-leading nations, where policymakers and industry leaders argue that slowing AI progress could allow strategic rivals, particularly China, to gain an advantage in global technological competition.
The White House has previously expressed concern over AI risks but has also emphasized the importance of maintaining US leadership in the sector.
Despite criticism, Anthropic said it plans to engage governments, scientists, and industry stakeholders in discussions aimed at establishing potential safeguards for frontier AI development.
The debate adds to growing global tension over how to regulate rapidly advancing artificial intelligence technologies while balancing innovation, competition, and safety concerns.
