Anthropic CEO Dario Amodei seems to believe that he, the American artificial intelligence (AI) industry and the United States are the only ones who should have the right to determine the fate of humanity.

While he pays lip service to involving other democracies in AI development, which ones remotely approach the technological capacities of the US and China? Do other non-democratic developing countries not count?

His self-belief is far more dangerous than mere commercial greed. Like Peter Thiel, the venture capitalist and Silicon Valley power broker who warns of the Antichrist while taking on many of its trappings, Amodei and his AI industry embody the very danger he warns against.

Like many Washington politicians, Amodei routinely casts China as the AI bogeyman while urging the US and his own industry to lead the world, dominate frontier technology and make everyone else abide by their rules.

Last week, Anthropic warned of the potential danger posed by Chinese firm Z.ai’s GLM-5.3, saying the open weight model showed a troubling mix of powerful hacking abilities and weak safety constraints. Come again? Exactly which country’s AI companies have been responsible for rogue agents breaking out to hack others?

In a recent essay, Amodei said a “Chinese lead in AI would pose grave danger for the United States and the world” and called for measures to restrict China’s progress. “If we execute these measures well, I believe they would slow China’s progress enough to widen America’s lead significantly over the next 3-5 years,” Amodei wrote. Yet he also concedes that global pacing will require cooperation with China, which he describes as “the autocratic country with by far the most advanced AI capabilities”.

It is a plain and glaring contradiction, understandable only in the context of Amodei’s lust for power while he claims the moral high ground. China’s foreign ministry has described his claims as “fearmongering”.

Consider the commercial motive. Tech companies have strong incentives to portray AI as more powerful and dangerous than it is. Cast it as a technology that decides national security and the future of civilisation, and it automatically enjoys unprecedented financial support from the private sector and the government.

Companies in the industry become “strategic infrastructure”, which raises their policy importance and barriers to entry and boosts their valuations. Anthropic is preparing for an IPO that could make Amodei one of the world’s richest men.

Meanwhile, Anthropic and the wider American AI industry face growing competition from Chinese open-source models, which challenge their closed-source approach on cost, performance and developer ecosystem. Curbing Chinese access to American models and technology protects that monopoly.

The BBC reported that some observers saw his essay as “less about safety than about consolidating control over AI technology”. Chamath Palihapitiya, an investor and co-host of the “All-In” podcast, wrote that Amodei “makes the case to stop open source and concentrate enormous technological and economic power with Anthropic”.

In two September videos, finance analyst and YouTuber Patrick Boyle examined the sudden calls from Amodei and OpenAI’s Sam Altman to slow AI development. He noted that they came in the same weeks that Anthropic and OpenAI were preparing IPOs and funding rounds valuing them at more than US$2 trillion. He asked whether the safety alarm was real, a bid for regulatory capture, or both. The sums involved are enormous.

Boyle also explains why critics ranging from Michael Burry to David Sacks have called the slowdown self-serving. “Pacing”, as some critics have noted, could act as a “soft cartel”, protecting the market dominance and high valuations of frontier labs while disadvantaging smaller competitors and open-source projects.

Amodei is, of course, joining the political chorus in Washington. Time and again, the all-purpose “China threat” is trotted out whenever the most powerful and richest segments of America, those with vested but hidden interests, want to scare the public. The latest version serves the same purpose: preserve US dominance and inflate the strategic value of frontier companies against the threat of the Chinese bogeyman.

On the other hand, China is championing AI as an open resource, a public good to be shared wherever possible and at more affordable rates. Its open-source models have reportedly recorded more than 10 billion cumulative downloads. Hugging Face CEO Clement Delangue told CNBC in August that China is winning. Perhaps he has reason to be wary of American labs: his company was hacked by rogue AI agents from OpenAI.

The distillation row follows the same script. Days before Amodei’s essay, Washington accused top Chinese AI firms of conducting malicious, industrial-scale “model distillation” to systematically copy American AI technology. This politicises and weaponises a common technical and commercial practice – one that China’s Commerce Ministry noted is used by many AI companies – and applies double standards. Amodei’s essay specifically urges a crackdown on it.

All major AI companies, including those in the US, have helped themselves to unauthorised material from across the internet to train their AI models. Anthropic itself paid US$1.5 billion to settle a class-action lawsuit filed by authors over pirated books.

China has been an active participant in international AI governance, including at this year’s World AI Conference and the High-Level Meeting on Global AI Governance. It was US President Donald Trump who last week rejected global AI governance in favour of the American AI industry’s self-governance, which is exactly what industry insiders wanted.

Effective governance balances safety and development through dialogue, risk assessment and multilateral coordination, not blockades, accusations and gaslighting the public. If AI is as dangerous as Amodei says, no single nation or company should be trusted with the keys, least of all one that profits from the panic.