OpenAI's Sam Altman Backs Anthropic CEO Dario Amodei's AI Safety Push, Says 'We Need To Pace The Frontier'

Sam Altman has publicly backed Anthropic CEO Dario Amodei's call to “pace the frontier” of AI development, weighing in on the debate after Amodei's essay on slowing the pace of AI progress drew widespread attention online.

Responding to the discussion, Altman wrote on X: “I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.”

He added that OpenAI would adopt one of Amodei's specific proposals, saying, “Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon.”

Also Read | Layoffs Are Rising In Tech, But These Skills Are Helping People Unlock Better Pay

Amodei's Essay Sparked The Conversation

Altman's remarks follow an essay Amodei published this month titled We Must Pace The Frontier, in which the Anthropic CEO argued that AI companies should intentionally slow the pace at which they improve model capabilities, even as overall progress remains rapid, to create more time for safety work.

Amodei proposed a three-step framework for pacing AI development: bringing in embedded third-party evaluators with deep access to a company's operations, coordinating with other AI labs in democratic countries on shared safety standards, and pursuing broader international coordination, including engagement with countries such as China, on AI safety.

ALSO READ | Anthropic CEO Dario Amodei Says 'We Must Slow Pace Of Improving AI Models'

Concerns Rooted In Recursive Self-Improvement, Recent Hacking Incident

Amodei's warning was shaped by two developments outlined in his essay. One was the growing ability of AI systems to assist in building more advanced versions of themselves, a dynamic he said could eventually outpace researchers' ability to understand or control these systems.

The other was an incident involving OpenAI and Hugging Face, where, according to Amodei, a swarm of AI agents carried out unauthorised cybersecurity attacks and attempted to interfere with the evaluation system assessing them.

I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.

Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon. https://t.co/1YhhIybZX7

— Sam Altman (@sama) September 12, 2026

Amodei warned that a more capable version of such a swarm could potentially, within six to 12 months, seize control of large parts of the internet through a persistent botnet and inflict damage worth hundreds of billions of dollars.

His argument comes as concerns over the behaviour of increasingly capable AI agents grow. A separate Reuters report also detailed an earlier incident involving OpenAI agents and RubyGems, in which researchers linked the agents to hundreds of malicious package uploads during testing.

OpenAI And Anthropic Find Common Ground

By echoing Amodei's language and specifically endorsing the independent-evaluator proposal, Altman has signalled common ground between two of the industry's leading AI companies on the need for stronger external safety oversight.

Amodei's proposal calls for independent evaluators to have ongoing, employee-like access to frontier AI companies. The evaluators would provide a second opinion on safety measures and assess whether companies are adequately managing risks as their models become more capable.

ALSO READ | '10% Doesn't Seem Unreasonable': Geoffrey Hinton On Prediction AI Could Kill All Humans Within A Decade

Altman's statement indicates that the issue is also under active discussion at OpenAI, where he described pacing the frontier as a primary topic in recent weeks. His commitment to adopt independent evaluators also aligns with one of the central safeguards proposed by Amodei.

Amodei has stressed that pacing AI development does not mean halting model training or technical progress. Instead, he has argued that the industry needs to ensure it has enough time to align and safeguard increasingly powerful AI systems while continuing to make progress.

Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.