Anthropic CEO Dario Amodei has called for the artificial intelligence industry to slow the pace at which it develops increasingly capable models, arguing that safety measures need more time to keep up with advances in the technology.
In an essay titled We Must Pace the Frontier, Amodei said AI development should continue, but at a pace that allows researchers, companies and governments to better understand and address emerging risks. His concerns include the potential loss of control over advanced AI systems, misuse for cyberattacks and biological threats, and wider economic disruption.
The warning comes amid growing debate within the AI industry over systems that could eventually improve their own capabilities with limited human involvement. This concept, known as recursive self-improvement, has raised concerns that increasingly autonomous systems could advance faster than researchers can understand or control them.
Amodei said the industry should focus on slowing capability development sufficiently for safety research and oversight to catch up. He proposed a three-part approach that includes independent monitoring of AI systems, coordination on safety standards among companies and democratic governments, and eventually broader international cooperation.
As part of the first step, Anthropic has committed to giving third-party evaluators ongoing access to its systems. These evaluators would be able to assess models during development, monitor adherence to safety measures and report potential incidents.
Amodei also suggested that AI companies should work together on common safety standards, while governments could establish regulatory frameworks for increasingly powerful systems. He acknowledged that effective oversight would eventually require international coordination, including discussions with countries that may have different political and technological priorities.
His comments have drawn support from other prominent figures in the AI industry. OpenAI CEO Sam Altman has backed the need for stronger external safety evaluation, while Elon Musk has also expressed support for slowing AI development to address potential risks.
However, calls for slowing the AI race remain complicated by commercial and geopolitical competition. AI companies are investing heavily in more capable models, while governments increasingly view artificial intelligence as strategically important for economic growth and national security.
Amodei has previously highlighted AI's potential benefits, including accelerating medical research, improving productivity and contributing to economic growth. His latest argument does not call for stopping AI development altogether, but for controlling its pace as capabilities increase.
The debate reflects a broader challenge facing the AI industry: how to continue developing increasingly powerful systems while ensuring that safeguards, monitoring and regulation advance alongside them.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.