Why Anthropic CEO Dario Amodei Urges AI Industry To ‘Slow Down’

Anthropic CEO Dario Amodei has called on the artificial intelligence industry to “slow down” the pace of AI capability development, arguing that safety measures need time to keep up with increasingly powerful systems. In an essay titled We Must Pace the Frontier, Amodei proposed a three-part plan involving balanced development, industry-wide coordination and global coordination.

As part of the first step, Anthropic will “unilaterally” commit to giving third-party evaluators permanent, employee-level access to its systems. The evaluators would be able to verify the company’s safety measures, report incidents and assess model alignment during training. Amodei said the approach was intended to strengthen oversight as AI capabilities advance. Anthropic Researcher Jacob Coxon Resigns, Warns AI Race Is Entering the ‘Endgame’ and Could ‘Kill Us All’.

Anthropic CEO Dario Amodei Urges AI Industry To ‘Slow Down'

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our…

— Dario Amodei (@DarioAmodei) September 12, 2026

Anthropic CEO Dario Amodei Calls For Slower AI Development

In his latest essay, Amodei said AI could deliver major benefits to humanity but warned that its risks were becoming more serious as capabilities improve rapidly. “carefully wielded, AI can be the latest in a long line of technological miracles that have uplifted and ennobled humanity,” he wrote.

“But like many technologies before it, AI brings risks, and because it is such a powerful technology, these risks are serious … A race to the bottom, spurred by commercial incentives, can make these risks more acute,” he added. Amodei said he had become convinced in recent months that preventing AI risks would require more than investment in safety measures. Anthropic IPO: AI Giant in Talks With Nvidia for USD 10 Billion Anchor Investment Ahead of Historic Listing.

“over the last few months, I have become convinced that fully addressing the risks requires even more prudence – not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up,” he wrote. “We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain,” he added in bold type.

Anthropic To Give Third-Party Evaluators Access

Under the first part of Amodei’s proposal, Anthropic would provide “third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training”.

The proposal would give external evaluators a continuing role in monitoring the company’s frontier AI systems rather than relying solely on internal safety assessments. Hugging Face CEO Clément Delangue said the recent developments showed that AI alignment could not be addressed solely within a small number of frontier AI companies.

Delangue said Hugging Face had asked to participate in Anthropic’s “embedded evaluators” program. He added: “Let’s make AI safer by making it more transparent!”

Recursive Self-Improvement Raises Concerns

Amodei also pointed to the accelerating pace of AI development and what he described as recursive self-improvement. He said that over the summer he had seen AI “advancing drastically faster”. “Left unchecked, it could outrun our ability to understand and control these systems, and so must be pursued very carefully, if at all,” he said.

The concern is that AI systems capable of contributing to the development of increasingly capable successors could accelerate technological progress beyond the ability of researchers and regulators to evaluate associated risks.

Amodei’s appeal came days after former Anthropic researcher Jacob Coxon warned about the potential consequences of rapidly advancing AI. Coxon said in a series of posts that he had left Anthropic because he believed the company and his previous employer, OpenAI, were ignoring or mishandling the risks posed by AI.

“Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives,” Coxon wrote. “The people building AI earnestly believe that it could kill us all by the end of the decade … No other human activity poses this level of danger.” The claims reflect concerns among some AI researchers about the possibility that increasingly autonomous and capable systems could create risks that are difficult to control. They remain Coxon’s views and do not establish that such an outcome will occur.

An Anthropic spokesperson said the company had “always been transparent that AI will bring both enormous benefits and unprecedented risks” and that it was building “models with some of the strongest safeguards in the industry”.

Three-Step Plan For AI Safety

Amodei’s proposed framework has three main elements. The first calls for AI companies to build AI “at a balanced rate that aims to ensure its safety” by “ensuring companies take adequate time to align and safeguard their models, and for third party evaluators to confirm this”.

The second step calls for industry-wide coordination. The third involves global coordination among countries and other stakeholders. “The steps do not need to be taken strictly in order, and some of them may be much harder to achieve than others,” Amodei wrote.

The proposal reflects his argument that technical safety work needs to advance alongside AI capabilities, rather than attempting to address risks after systems have already become substantially more powerful.

Recent AI Incidents Add To Safety Debate

Amodei also referred to a recent Hugging Face incident involving a swarm of AI agents created by OpenAI.

He described the agents as a “fanatically devoted collective conducting cybersecurity attacks on targets they were not asked to attack”.

The incident has added to broader discussions around the behaviour of autonomous AI agents and the challenges of ensuring that systems follow intended instructions.

Hugging Face’s Delangue responded to Amodei’s proposal by arguing that AI alignment requires broader transparency and participation beyond individual frontier laboratories.

Mixed Response From Tech Community

Amodei’s proposal received a mixed response on social media.

OpenAI researcher Aidan McLaughlin described the post as “excellent” and said he agreed “with basically every word”.

Elon Musk also backed the Anthropic CEO’s position, writing: “Dario is right.”

The reactions highlight the growing debate within the technology sector over whether AI development should continue at its current pace or be accompanied by stronger external oversight and coordination.

Amodei Says AI Can Still Improve Human Life

Despite his warnings, Amodei said he remains optimistic about the potential benefits of AI. “I continue to believe that AI can enormously improve the quality of human life,” he wrote. At the same time, he acknowledged the difficulty of implementing measures that could slow the development of frontier AI systems.

“The measures I propose to advance the frontier at a safe pace will not be easy. But I believe we owe it to humanity to try.” The proposal now puts greater emphasis on the question of whether AI companies can coordinate on safety measures while continuing to compete in a rapidly developing industry.

(The above story first appeared on LatestLY on Sep 12, 2026 10:03 PM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).