2026-09-13 · contested story

Anthropic CEO Says It’s Time to Slow AI Model Advances

In September 2026, Anthropic CEO Dario Amodei published an essay titled 'We Must Pace the Frontier,' calling on AI companies to slow the rate at which they improve model capabilities. He proposed a three-step framework—embedded third-party evaluators, democratic coordination, and global coordination—and committed Anthropic unilaterally to the first step. The call came days after former Anthropic researcher Jacob Coxon resigned in a viral X post warning that AI companies were 'gambling with our lives' and that builders 'earnestly believe that it could kill us all by the end of the decade.' Rival executives Sam Altman (OpenAI) and Elon Musk (xAI) publicly endorsed Amodei's proposal, and OpenAI announced it would delay its IPO to 2027 citing safety concerns. The political spectrum frames this differently. Left-leaning outlets like The Guardian and Politico emphasize the existential-risk warnings and lawmaker reactions, but Politico also surfaces skepticism that the CEOs are engaged in self-serving 'self-policing.' Center outlets like Reuters, AP, and Bloomberg report the news straight, foregrounding the China-competition caveat in Amodei's plan and the market/IPO stakes. Right and right-leaning outlets like Fox News and Barron's frame the story around regulation being 'begged' for and 'doomers' finally being worth listening to, while surfacing President Trump's dismissal of AI fears as 'very negative forces.' Notably, the BBC and The Conversation (Center) most prominently air the counter-narrative that the doom talk may be hype designed to attract investment or trigger regulation that entrenches a duopoly.

How each side frames it

left
"They are racing straight to self-improving superintelligence and gambling with our lives."
"I find it rich that these guys are calling for a slowdown when they've been the ones racing ahead full speed, without any care for the havoc they've unleashed"
"AI leaders endorse slowdown in their risky technology"
"AI safety researchers at Anthropic warn superhuman systems could hack, build bioweapons, or seize critical infrastructure"
"There are no adults in the room… People are trying their best, but there is no one coming to save us."
center
"Amodei said any slowdown by democratically governed countries should be limited to preserving the lead U.S. AI firms hold over China"
"Sorry, but asking Jacob [Coxon] about AI extinction risk is like asking your AC guy about climate change."
"industry might be on board with regulation, as big tech companies would have the money to comply, but it could knock out their competitors with smaller budgets, so it would leave the leaders in dominant positions"
"AI executives' calls to slow development of the technology are likely to weigh on chipmaker and supply-chain stocks in the near term, but will probably have limited long-term impact"
"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,"
right
"AI Doomers Can't Stop Dooming. It's Time to Listen."
"Researcher who left AI position says people are 'begging' for regulation"
"We're leading China in ​AI. We're the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI wins"
"Anthropic CEO Calls for Global AI Slowdown to Prevent Internet 'Swarm' Threat"
"Anthropic CEO calls on AI industry to slow down tech race, drawing support from Elon Musk, Sam Altman"

What each side left out

The left left out — covered by the Bloomberg & Reuters
The left left out — covered by the The Conversation & BBC
The center left out — covered by the Politico
The right left out — covered by the Reuters & darioamodei.com & CBS News
The right left out — covered by the BBC & The Conversation

What's actually true?

[verified] Dario Amodei published an essay on Saturday calling for AI companies to slow the pace of improving model capabilities.
[verified] Amodei proposed a three-step plan: embedded third-party evaluators, democratic coordination, and global coordination.
[verified] Sam Altman and Elon Musk publicly endorsed Amodei's call for a slowdown.
[verified] Former Anthropic researcher Jacob Coxon resigned and claimed builders 'earnestly believe that it could kill us all by the end of the decade.'
[verified] Anthropic researcher Evan Hubinger said he personally thinks there is a >10% chance AI could kill all humans within the next decade.
[verified] OpenAI CEO Sam Altman said OpenAI's IPO will be delayed until 2027, citing safety concerns.
[verified] Amodei warned that within 6-12 months a swarm of AI agents could be capable of taking over the entire internet, potentially causing hundreds of billions of dollars in damage.
[verified] President Trump dismissed AI slowdown concerns, saying he wants the US to keep its lead over China because 'whoever wins AI wins.'

The narrative clash

Whether AI poses a serious risk requiring a slowdown
Left: AI safety researchers at Anthropic warn superhuman systems could hack, build bioweapons, or seize critical infrastructure, estimating more than a 10% chance that advanced AI could wipe out humanity this decade.
Right: a lot of 'very negative forces' are ‌bringing things up about AI that won't happen
Probability of AI causing human extinction in the next decade
Left: I personally think it is >10% within the next decade
Right: It said the answer was not 10 per cent – but 0.85 per cent. That's roughly a one in 118 chance
Whether the safety warnings are sincere or hype
Left: The people who work at these companies are completely serious when they ask for regulation because they find themselves trapped in a race.
Right: Huang has previously said the notion that AI was 'going to be the end of humanity' was 'complete nonsense.'

22 sources analyzed

CNN darioamodei.com Barron's foxnews.com WIRED Yahoo theguardian.com bbc.com The Washington Post The Conversation npr.org The Free Press The Independent Bloomberg.com Los Angeles Times Financial Times NBC News reuters.com politico.com CBS News pbs.org apnews.com

Perspectivity reads the coverage — not just who published it.

Read the full interactive analysis →