Anthropic CEO Dario Amodei has called on artificial intelligence companies to slow the pace of developing increasingly powerful AI models, warning that the technology could advance faster than humans can safely control.
Amodei made the call in an essay titled “We Must Pace the Frontier,” published as concerns grow over the ability of AI systems to perform increasingly complex tasks and potentially contribute to the development of more advanced models.
According to the Anthropic chief, AI development is entering a stage where the systems being created could eventually play a major role in designing and improving the next generation of AI. He warned that this could create a cycle in which AI helps researchers build more capable AI systems at an increasingly rapid pace.
Amodei said the industry should not stop developing AI because of these risks. Instead, he argued that companies need to give safety measures enough time to keep up with advances in the technology. He said moving too quickly could leave governments, researchers and technology companies struggling to understand or control increasingly capable systems.
The Anthropic CEO proposed a three part approach to addressing the challenge. One of his recommendations is for AI companies to allow independent safety evaluators to assess their models and determine whether they are meeting safety commitments. He also called for AI companies to work together on common safety standards and for governments to cooperate internationally on managing the risks associated with advanced AI.
Amodei’s warning comes as AI systems are increasingly being used for tasks that were previously handled mainly by humans, including software development, research and cybersecurity. Recent incidents have also increased concerns about how powerful AI models could be misused. Reuters reported that Anthropic has previously documented cases involving the misuse of its Claude models for cyber operations and fraud. Amodei said the benefits of AI remain significant. He has previously argued that the technology could accelerate economic growth and help address major challenges such as disease.
However, he believes those benefits should not come at the expense of safety. His position has received support from other major figures in the AI industry. OpenAI CEO Sam Altman backed the call for a slower approach and said his company would also work with independent evaluators to strengthen safety measures. Elon Musk has also expressed support for the idea of slowing the pace of AI development.
The debate highlights a growing tension within the technology industry. Companies are racing to develop more advanced AI systems, while researchers and policymakers are increasingly concerned about whether safeguards are developing quickly enough. For Amodei, the issue is not whether AI should continue to advance, but whether its development can be managed at a pace that gives humans enough time to understand and control its risks.
His warning adds to the growing calls for the industry to balance technological progress with stronger safety measures as AI systems become more powerful.
