Anthropic claims AI is too fast and needs to hit the brakes: let’s take it with a pinch of salt

Anthropic, the company behind the Claude chatbot, has never really advocated for any pauses in the AI race but has suddenly changed its tune, warning that frontier models could become increasingly difficult for humans to control. Skeptics say it’s all part of the hype, though.

A global pause on building the most powerful AI systems is needed, Anthropic said on Thursday, as the latest models are beginning to show signs they could escape human control.

“We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology,” said the San Francisco-based company.

Wonders and dangers

Typically, the post on Anthropic’s website is extremely long and full of braggado about the capabilities of Claude, the firm’s chatbot.

The numbers are certainly interesting, but difficult to verify as they’re produced by Anthropic itself.

For instance, the company claims that AI is already dramatically speeding up the development of AI itself and warns that this could eventually lead to “recursive self-improvement.”

“As of May 2026, more than 80% of the code we merge into Anthropic’s codebase was authored by Claude,” said Anthropic before, to be fair, admitting that lines of code are an imperfect measure since it values quantity over quality.

An Anthropic employee – anonymous – is also quoted as saying: “I started leaning hard into Claudifying about a year ago. That’s been a crazy adventure, and it’s now been ~5 months since I last wrote any code myself.”

We wonder if they still have a job at the company, of course. Nevertheless, Anthropic is really persistent.

“AI that can build itself would be a major development in the history of technology – one that could bring enormous good for the world in science, healthcare, and beyond,” the post claims.

“But full recursive self-improvement also might increase the risks of humans losing control over AI systems.”

Is meaningful coordination even possible?

However, all this sounds a little overdramatic, skeptics say. Critics of the AI industry have long said that the doom talk is actually a strategy for slowing rivals under the cover of safety concerns.

Plus, a coordinated slowdown of AI development is just not going to happen. Even if one player stops, others – rival companies or countries like China – will keep playing.

US officials and tech executives have indeed repeatedly argued that any slowdown in AI development risks handing China a crucial strategic edge – even if US President Donald Trump signed an executive order this week that allows the government 30 days to conduct a preliminary review of the most powerful US AI models before their release.

To be fair, Anthropic seems to recognize the naivety of its proposal, comparing the problem to nuclear arms control treaties and noting that AI training is far easier to hide than a missile silo.

“A meaningful slowdown or pause would require multiple well-resourced labs at or near the frontier, in multiple countries, agreeing to stop under the same conditions. It would also require that each can verify that the others have actually stopped,” the company admits.

Anthropic now plans to bring together government officials, scientists, advocacy groups, and rival AI firms in the coming months to figure out how to move forward.

Has AI lowered the costs of anything, actually?

So why now? After all, Elon Musk and other big thinkers already signed a “Pause Giant AI Experiments” open letter back in March 2023, also urging a slowdown in AI development.

One theory claims that Anthropic, like other AI giants such as OpenAI, wants to push the brakes because it's simply running out of GPU capacity. Developers everywhere are burning too many tokens left and right.

Another suggests that Anthropic is reorganizing the hype because it’s preparing to go public. The company, valued at nearly $1 trillion, has indeed filed for IPO with the Securities and Exchange Commission.

The move sets up an early test of whether investor appetite for AI will hold up under public scrutiny. So far, the astronomical valuations have all been private.

Some observers, though, claim Anthropic is actually admitting that AI isn’t some magic and point to this paragraph from the company’s post:

“Achieving recursive improvement alone does not suggest an immediate change in how industrial production occurs, societies organize, or markets function. More intelligence can’t learn what a drug does over decades of use, can’t hold elections sooner than a constitution dictates, and can’t turn a stranger into an old friend in a weekend.”

According to prominent tech creator and software engineer Mo Bitar, Anthropic is questioning whether AI may turn out to be altogether useless, and this is “the single most honest thing Anthropic has ever written.”

“Anthropic’s only measure for AI productivity is lines of code shipped. Meanwhile, has AI lowered the cost of healthcare? Has it made food or rent more affordable? Has it made college more affordable?” said Bitar.

Unlock more exclusive Cybernews content on YouTube.