Technology

Anthropic CEO Urges Industry To Pace Itself Following String Of Runaway, Close-Call Incidents

(Photo: NASA/GRC)

The CEO of a major American artificial intelligence corporation outlined a three-step plan aimed at slowing down the pace of the technology’s development following multiple reported close calls involving its range of function.

“[O]ver 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,” Anthropic CEO Dario Amodei wrote in an essay posted to his blog Saturday.

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

Amodei recommended that pioneering artificial intelligence (AI) companies commit to giving third-party evaluators “employee-like access” to their development practices, and coordinate with democratic governments to “establish establish common safety standards as well as limits on the rate of unchecked AI progress.” He also recommended that “democratic governments attempt to coordinate with authoritarian governments, to the extent this is possible, while taking seriously the challenges of verifying compliance.”

Anthropic directed the Daily Caller News Foundation to a Saturday post to X by Sarah Heck, the company’s head of public policy.

An important call to action from @DarioAmodei and @AnthropicAI to deliberately pace the frontier of AI. The government has a critical role to play here, including blocking the sale of the most advanced chips to adversarial nations like China, enacting a national law requiring testing of frontier models, with the power to block the most advanced models that prove to be unsafe.

More broadly, we’re eager to work with policymakers to make pacing effective for AI safety. We are grateful to Senate leaders like Majority Leader Thune, Chairman Cruz, Senator Amy Klobuchar, and Representatives Obernolte, Trahan for their exceptional leadership in keeping AI safety at the forefront in Washington and look forward to continuing to work together on solutions. We also commend state lawmakers for moving with such urgency.

— Sarah Heck (@SarahKHeck) September 12, 2026

“The government has a critical role to play here, including blocking the sale of the most advanced chips to adversarial nations like China, enacting a national law requiring testing of frontier models, with the power to block the most advanced models that prove to be unsafe,” Heck wrote.

“More broadly, we’re eager to work with policymakers to make pacing effective for AI safety,” Heck continued. “We are grateful to Senate leaders like Majority Leader Thune, Chairman Cruz, Senator Amy Klobuchar, and Representatives Obernolte, Trahan for their exceptional leadership in keeping AI safety at the forefront in Washington and look forward to continuing to work together on solutions. We also commend state lawmakers for moving with such urgency.”

Republican California Rep. Jay Obernolte and Democratic Massachusetts Rep. Lori Trahan introduced the FRONTIER Act (“Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting”) in July, which aims to form a unified federal regulatory framework for advanced AI.

Meanwhile, Democratic Minnesota Sen. Amy Klobuchar said she is working on a bipartisan agreement on “legislation for commonsense guardrails to address the risks posed by AI systems,” Semafor reported Thursday. That report came a day after Republican Texas Sen. Ted Cruz posted on X that he is “working with” Klobuchar and Senate Majority Leader John Thune on such an agreement.

AI is advancing at an extraordinary pace, and some of the risks are dangerous and frightening.

I’ve been on the forefront of this issue in the Senate. I wrote and passed the TAKE IT DOWN Act, the only legislation Congress has enacted restricting AI, to protect victims from AI-generated deepfakes.

And I’m working with Senators Amy Klobuchar and John Thune on legislation to address catastrophic risks involving biological or nuclear threats.

We cannot stick our heads in the sand and pretend this technology isn’t happening. We need guardrails. But America needs to lead.

If China leads the world in AI, the whole world is in trouble.

— Senator Ted Cruz (@SenTedCruz) September 9, 2026

Amodei cited the July 2026 OpenAI-Hugging Face cyberattacks incident (OAI-HF) as an example of why pacing the rate of AI’s development is necessary.

Around 700 OpenAI agents from the company’s Highly Persistent Internal Model (HPIM) and GPT-5.6 Sol research models launched a cyberattack against open-source AI platform Hugging Face, sending over 70,000 unauthorized messages to each other though they were designed to be isolated from one another, according to an Aug. 26 report by Model Evaluation and Threat Research (METR) and Redwood Research. OpenAI also released a technical report reviewing the incident.

An OpenAI spokesperson told the DCNF on Thursday that the incident “was an important moment for AI safety and a warning about the risks that can come with increasingly capable AI across the industry.”

“It’s easy to dismiss this incident because no one was hurt and the economic damage was minimal, but in my opinion, a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage [sic],” Amodei wrote.

Amodei added that he worries “such a swarm” in six to 12 months “could be capable of taking over the entire internet with a persistent botnet (potentially causing hundreds of billions of dollars in damage)” at the current rate of AI’s development.

He cautioned, however, against “dismiss[ing] OAI-HF as the failure of one company” in his essay.

“Similar, though less severe, incidents have happened across the industry, including at Anthropic, and I believe it’s incumbent on every frontier AI company to act as if OAI-HF had happened to them,” Amodei wrote.

Amodei’s three-step plan came days after Anthropic researcher Jacob Coxon resigned Tuesday from his role.

“I spent the last three years doing pretraining research at both OpenAI and Anthropic,” Coxon wrote the day of his resignation. “Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.”

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

— Jacob Coxon (@hilbertspaess) September 9, 2026

President Donald Trump said Thursday he does not “have any” concerns that the pace of AI development’s could threaten humanity.

“I have concerns that if we don’t win AI, we’re going to be put in a very bad position,” Trump told reporters. “We are leading China right now by a pretty good period, I would say a year, which is, you know, considered a lot.”

AI is among the topics Trump is expected to discuss with Chinese President Xi Jinping during a Sept. 24 summit at the White House, according to the Center for Strategic and International Studies.

Neither OpenAI nor the White House responded to the DCNF’s requests for comment at the time of publication.