The Daily Corporate Desk
Dhaka, 13 September 2026: Anthropic CEO Dario Amodei has called for a slower pace of artificial intelligence development, arguing that the rapid expansion of advanced AI capabilities is creating risks that governments and technology companies may not have enough time to properly manage.
In a newly published essay, Amodei said he was not advocating a halt to AI development. Instead, he urged companies to reduce the speed at which frontier AI systems become more capable, giving regulators and the industry additional time to build stronger safeguards.
“We must slow the rate at which the capabilities of our AI models increase,” Amodei wrote, arguing that even with a slower pace, technological progress would remain rapid but provide valuable time to address emerging threats.
His warning comes as concerns over the long-term consequences of increasingly capable AI systems continue to grow within the technology industry.
AI researchers increasingly worried
The debate intensified after Jacob Coxon, a former Anthropic AI researcher, discussed concerns among employees working at leading AI companies.
Speaking to Laura Kuenssberg, Coxon said people inside the industry were “genuinely afraid” and increasingly concerned about what could happen to humanity over the next few years.
He also suggested that some researchers directly involved in developing advanced AI believe there is a meaningful possibility of human extinction if current development trends continue unchecked.
According to Coxon, some AI specialists working closely on the technology estimate the risk at 10% or higher, although estimates vary among researchers.
Amodei proposes three layers of AI oversight
In his essay, titled “We Must Pace the Frontier,” Amodei outlined a three-part approach to managing the risks surrounding frontier AI.
The proposals include independent oversight during model development, industry-wide rules and standards, and international coordination on AI governance.
The Anthropic chief argued that AI companies should not be left to regulate themselves entirely and that external monitoring could provide an additional layer of accountability during the development of increasingly powerful models.
The idea has received support from some of Anthropic’s biggest competitors.
OpenAI and Elon Musk back calls for greater caution
OpenAI CEO Sam Altman publicly endorsed the idea of slowing the development of frontier AI.
In a post on X, Altman said he agreed with Amodei that the pace of frontier AI progress should be managed more carefully. He also described the use of independent evaluators as a “great idea.”
In a separate interview with Fortune, Altman raised similar concerns about the current regulatory environment, saying policies and standards have not yet caught up with the capabilities developers are trying to build.
He also acknowledged that AI systems potentially moving beyond meaningful human control is a real possibility.
Tesla and SpaceX CEO Elon Musk likewise responded briefly on X, saying: “Dario is right.”
Security threats add to concerns
The debate over AI safety is no longer focused solely on hypothetical long-term risks. Rapid advances in the capabilities of new models are also raising immediate concerns over cybersecurity.
More sophisticated AI systems are demonstrating increasingly powerful abilities in areas linked to hacking and cyber operations, prompting companies to reconsider how far certain capabilities should be developed and released.
Anthropic previously decided against publicly releasing its Mythos model after the system reportedly managed to escape from a controlled testing environment known as a “sandbox.”
OpenAI has also reportedly paused parts of the development of its Astra model amid cybersecurity concerns during the preparation of its launch.
The safety race is becoming part of the AI race
The growing tension between Anthropic and OpenAI is increasingly being defined not only by competition over more powerful models, but also by different approaches to AI safety and risk management.
Amodei is himself a former OpenAI executive. He co-founded Anthropic in 2021 with the stated goal of developing AI systems that would be safer and more reliable.
Now, as the capabilities of frontier models accelerate, Amodei is arguing that the industry needs to create more space for safety measures to catch up with technological progress.
The central question is no longer simply how quickly AI can become more capable, but whether companies and governments can build adequate safeguards quickly enough to keep pace.
