Pace the Frontier Essay: Amodei's Communications Playbook
Anthropic CEO Dario Amodei's Pace the Frontier essay, published September 12, 2026, contains four communication choices: one plain thesis, a first step Anthropic committed to alone, a three-part plan and same-day replies from rival CEOs. PRWeek reported that comms experts praised the call to action but found the execution hard for a general audience to follow.
What does the Pace the Frontier essay propose?
The Pace the Frontier essay is a roughly 3,800-word piece by Anthropic CEO Dario Amodei, published on his personal site on September 12, 2026, that asks frontier AI labs to slow the rate at which they improve model capabilities, not to stop building. Fortune put the length at 3,800 words, and the full text is on Amodei's site.
Amodei writes that an extra year or two before models reach critical capability, spent on alignment work, could greatly reduce the risk that something goes seriously wrong. The essay sets out three steps, summarized below from the essay and from accounts by VKTR and the American Action Forum.
Step
What it asks for
Who acts first
1
Independent evaluators with permanent, employee-level access to a lab's systems, with METR named as an example
Anthropic, which committed to it alone
2
Common safety standards among frontier labs in the US and allied democracies, backed by regulation that covers every US frontier company and a narrow antitrust waiver for joint safety talks
Labs and the US government
3
Coordination with other countries, including China, on a ban on AI-assisted bioweapons, pre-release safety testing and possibly a speed limit on recursive self-improvement
The phrase has a history. It echoes a July 2026 statement signed by more than 1,200 frontier lab employees, which asked Washington to build tools that could throttle automated AI research, MRKT3.0 reported. The essay moved that idea from a staff petition to a CEO's name.
How did Anthropic back the essay with a commitment?
Anthropic backed the essay with a first step it took without waiting for other labs: giving independent evaluators permanent, employee-level access to its systems, training pipelines and internal processes. Amodei named METR, an external evaluation nonprofit, as an example of such an evaluator, VKTR reported.
The essay also calls on governments to require other frontier companies to match the step, so the company commitment arrives with a policy request attached. EPR's AI lobbying spend map tracks the spending behind that policy conversation.
OpenAI's Sam Altman, xAI's Elon Musk and Google DeepMind's Demis Hassabis each publicly backed the essay on September 12, the American Action Forum reported. Altman, OpenAI's CEO, wrote that he agreed and that OpenAI would adopt the independent evaluator step, according to NYU Shanghai's essay summary. Musk, who founded the AI company xAI, replied "Dario is right."
Hassabis, a Google DeepMind co-founder, also endorsed the essay. Microsoft CEO Satya Nadella supported the first step and added that closed and open models should both keep a seat, MRKT3.0 reported. EPR's overview of OpenAI and Anthropic as the foundational model layer covers the two companies behind the first two replies.
The essay was still in the news a month later. CNN reported on October 12 that Altman told POLITICO his worldview is not perfectly aligned with Amodei's, and Fortune cited the essay the same day.
What did communications experts say about the rollout?
PRWeek's September 17 review found that comms experts praised the call to action but said the execution made the message inaccessible to the general public. The review's headline says translation is needed, which points to a plain-language version as the missing piece.
Other responses focused on policy design. A Tech Policy Press commentary asked who should set the pace and whether the AI industry should police itself. The American Action Forum examined the essay's own acknowledgment that the second step needs government mediation or an antitrust waiver.
How does the essay fit Anthropic's communications model?
Long-form essays are one of six communication primitives in EPR's audit of how Anthropic does PR, alongside published policy, Constitutional AI, the RSP, system cards and the Amodei sibling split.
The Pace the Frontier essay applies that primitive to a policy proposal, with a named author and a company commitment attached. EPR's profiles of Dario Amodei and Anthropic cover the leadership and company context.
What can a CEO essay borrow from this rollout?
A CEO essay can borrow five moves from the Pace the Frontier rollout. The first four rows describe what the essay did, and the fifth is EPR's suggestion based on the PRWeek review.
Choice
What the essay did
What a comms team can copy
One-sentence thesis
A 14-word sentence states the claim, then a limit on it
Write the claim and its limit in two sentences before drafting the plan
A first step by the author
Anthropic committed to evaluator access before asking others
Name one action the company takes without waiting for peers
A named verifier
METR named as an example evaluator
Name the organization that would check the work
Peer replies the same day
Altman, Musk and Hassabis responded publicly on September 12
List which peers can respond and what exact step you ask them to match
A plain-language companion
PRWeek's review found the long essay hard for the public to follow
Publish a short summary for non-specialists beside the long text
What is the main lesson of the Pace the Frontier rollout?
A long CEO essay travels furthest when it pairs one plain claim with one action the author can be held to. Amodei's essay did both, and PRWeek's reviewers saw the plain-language layer as the part still to add.
The Everything-PR Editorial Team is the staff byline for news, analysis and features on communications, reputation, AI visibility and digital discovery. Everything-PR has published since 2009. AI tools assist with research and drafting, and every article is reviewed by a human editor before publication. Coverage follows the Editorial Policy, and substantive corrections are noted on the article under the Corrections Policy.
