AI CEOs back Anthropic plan to slow frontier development

Anthropic chief executive Dario Amodei proposed at the weekend that artificial intelligence companies slow the pace of frontier development, in an essay titled “We Must Pace the Frontier”. OpenAI chief executive Sam Altman, Google DeepMind chief Demis Hassabis and xAI owner Elon Musk publicly backed the proposal.

The proposal followed repeated warnings from Anthropic researchers that artificial intelligence could kill all of humanity by 2030, which drew a public uproar. An Anthropic researcher who quit said AI staff fear for humanity.

What does Amodei’s pacing proposal say?

Amodei’s proposal rests on three pillars: outside evaluators granted employee-level access to monitor risk; agreements between democratic states on AI safety standards; and global coordination with authoritarian governments such as China on security issues. Amodei and other industry figures have made similar calls for global coordination in the past.

Which AI executives backed the plan?

Altman posted on X on Saturday that he would match Amodei’s commitment to embedding outside evaluators inside his own company. “I agree with Dario that we need to pace the frontier,” Altman wrote. “Committing to having independent evaluators with employee-like access is a great idea, and we will do the same.” Hassabis wrote: “Dario’s essay points towards the right path forward.”

Anthropic and OpenAI did not immediately respond to a request for comment.

Why do researchers doubt the plan?

Critics allege that AI firms are trying to pre-empt more stringent government regulation and retain the status quo of power within the industry. Rahm Emanuel, a former congressman and chief of staff to Barack Obama, wrote on X that it was “an admission that we’re driving down a dark, winding road on wet pavement with the headlights off”. Sarah Myers West, co-executive director of the AI Now Institute, wrote: “AI exceptionalism has led to a total failure to uphold our standards for what products are deemed safe for use.”

Alex Bores, whose failed bid for Congress drew $24m in spending both for and against his campaign from AI-linked political action committees, warned that the AI lobby seeks to undermine legislation while claiming to support regulation. “To everyone who genuinely wants real AI regulation, please don’t fall for their games,” Bores posted on X on Sunday.

What is the antitrust objection?

Amodei’s proposal suggests that governments could grant some form of antitrust waiver or exemption to allow industry coordination. David Sacks, Trump’s former “AI czar”, and Alvaro Bedoya, the Biden-era head of the Federal Trade Commission, suggested the proposal amounted to regulatory capture that would benefit current industry leaders. Bedoya wrote on X that antitrust law “does not prevent AI companies from coordinating to make sure AI does not hurt people” but “does absolutely prevent AI companies from organizing to prevent the entry of cheaper, upstart rivals”.

What barriers remain to AI regulation?

The US federal government is committed to a global AI arms race, previous proposed AI legislation has wilted under intense, well-financed corporate lobbying, and AI startups face pressure to release new, more powerful products ahead of trillion-dollar public offerings. AI firms have also shifted how much they emphasise potential harm: OpenAI shut down a “superalignment” team dedicated to long-term AI risks in 2024 and disbanded another team intended to communicate how OpenAI would benefit humanity earlier this year.

Frequently asked questions

What is the “We Must Pace the Frontier” proposal?

It is a plan published at the weekend by Anthropic chief executive Dario Amodei calling for artificial intelligence companies to slow frontier development. It rests on outside evaluators with employee-level access, AI safety agreements between democratic states, and coordination with authoritarian governments such as China.

Which AI executives supported Dario Amodei’s pacing plan?

OpenAI chief executive Sam Altman, Google DeepMind chief Demis Hassabis and xAI owner Elon Musk posted support. Altman said he would match Amodei’s commitment to embedding independent evaluators with employee-like access inside his own company.

Why do critics oppose the AI pacing plan?

Critics including former Obama chief of staff Rahm Emanuel and former Federal Trade Commission head Alvaro Bedoya say the plan could pre-empt stricter government regulation and amounts to regulatory capture. Bedoya singled out the proposal’s suggestion that governments could grant an antitrust waiver for industry coordination.

What obstacles remain to AI regulation in the United States?

The US federal government is committed to a global AI arms race, previous proposed AI legislation has wilted under well-financed corporate lobbying, and AI startups face pressure to release more powerful products ahead of trillion-dollar public offerings.

With reporting from The Guardian.

Stock image, not from the events described. Global Map — World Travel Adventures / Wikimedia Commons (CC0 1.0)