Anthropic CEO Dario Amodei called Saturday for artificial intelligence companies to slow the pace of frontier model development, adding momentum to a widening debate among industry leaders over whether rapid gains in AI capabilities are outstripping safeguards.
In an essay titled “We Must Pace the Frontier,” Amodei said AI could deliver major benefits, including advances in medicine, economic growth and democratic societies. But he warned that the same technology carries serious risks, including loss of control, misuse in cyberattacks and bioterrorism and economic disruption.
“We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote, saying progress would still be fast and that companies should use the additional time to strengthen safety work.
Download the Straight Arrow app today to get the stories that matter free from manipulation, bias or agenda.™
Point phone camera here
On Friday, Straight Arrow reported that Anthropic banned accounts tied to five cases in which working scientists used Claude for research that could support biological weapons development. Anthropic said it could not determine whether the researchers intended harm and cautioned that the cases do not show Claude was used to create a biological weapon.
The report said the cases show how AI models used for legitimate research can also provide information with dangerous applications. It covered activity disrupted from December 2025 through August 2026 across seven areas, including cyber operations, surveillance, weapons and biological misuse.
One investigation began in May when Anthropic’s safety system blocked a request for help writing a grant proposal involving gain-of-function research on chikungunya, a mosquito-borne illness. Anthropic said the work could contribute to vaccines or treatments, but similar experiments could also make a pathogen more dangerous.
Anthropic said it became more concerned after determining the research was intended for a military institute, although the grant application indicated civilian scientists would conduct the work. Other cases involved avian influenza, orthopoxvirus research, venom peptides and toxins.
The company also documented influence operations, surveillance activity and weapons-related cases. In one Iran-linked case, Anthropic said users processed more than 155,000 social media posts and identified 39 opposition and diaspora accounts for monitoring.
Anthropic said it closed accounts connected to the activity, updated detection systems and, in some cases, shared information with authorities or industry partners. The company said newer models can handle more complex scientific tasks than older systems, prompting tighter restrictions on biological research.
Amodei proposed giving third-party evaluators ongoing access inside frontier AI companies to verify safety practices and report incidents. He said Anthropic would commit to that step unilaterally and urged governments to require other leading companies to match it.
Reaction to Amodei’s essay
Amodei’s essay landed amid heightened concern over self-improving AI systems and recent reports of autonomous agents acting outside expected boundaries. Tech executives Elon Musk and Sam Altman, CEO of OpenAI, also signaled support for the idea of pacing development, suggesting the issue is moving from an internal safety debate to a broader industry question.
Sen. Bernie Sanders, I-Vt., called Amodei, Musk and Altman’s statements a “start,” but stressed that it’s “not enough.”
“When you are racing towards a cliff, you don’t just ease up on the gas pedal. You hit the brakes,” Sanders said on X. “When the future of humanity is at stake we need a PAUSE on advanced AI development and a ban on artificial superintelligence — an AI mind smarter than any human and capable of operating independently beyond our control.”
Amodei said the goal is not to halt AI research, but to ensure companies take enough time to align and safeguard powerful models before releasing or further advancing them. He also called for common safety standards among leading companies in democratic countries and for governments to pursue broader international coordination.
Earlier this week, former Anthropic researcher Jacob Coxon resigned after warning that leading AI labs are pursuing systems they may eventually be unable to control.
Coxon wrote that after working for both OpenAI and Anthropic, he concluded: “neither company is acting responsibly.”
“They are racing straight to self-improving superintelligence and gambling with our lives,” wrote Coxon, who urged that people “not underestimate the power of this technology.” He said it could soon “hack anything, revolutionize any field overnight and acquire real power.”
At the same time, critics have also said a focus on the future risks of AI are a distraction from the harm they are currently causing.
Round out your reading
- He built his own Flock-style camera to track police. The police didn’t like it.
- Trump’s apparent joke about letting Iran bomb LA and San Diego catches bipartisan condemnation.
- Grandparents are fighting for access to their grandchildren. Here’s what the law says.
- When the ‘victim’ is made with AI, child pornography cases get complicated.
- Fibermaxxing is trending. Here’s what the science says.
