The AI pause divides opinion: Meta refuses to slow down, while Mistral denounces OpenAI’s and Anthropic’s strategy.

The AI world has been shaken by several calls to slow down in recent days. For Dario Amodei, CEO of Anthropic, it is imperative to pull back before autonomous agents can take control of the Internet.The leader highlighted a three-step plan to slow the pace of AI advances without hampering innovation. Sam Altman, OpenAI’s CEO, agreed with his rival, saying he would take strong measures to improve how its AI systems are evaluated.

Even Elon Musk echoed the same sentiment, followed a few days later by Bilal Chughtai, a former research engineer at Google DeepMind. The researcher is convinced “that AI has the potential to kill us all and that we may be running out of time to prevent such an outcome”. In this context, OpenAI, Google, and Anthropic began collaborating on a joint framework for overseeing the risks of artificial intelligence.

Also read: According to Anthropic, plumbers and electricians could be the big winners of AI

Mark Zuckerberg refuses to back down on AI

The call to slow down issued by Anthropic and OpenAI was not welcomed by all AI giants. Mark Zuckerberg, the head of Meta, firmly rejected the idea of a slowdown in research. He believes that natural safeguards already exist that can prevent AI from becoming a danger. According to him, “people will not want to use AI agents that are not aligned with their interests and that do not do what they ask”.

In fact, “laboratories therefore have a strong natural incentive to make their models more aligned”.In plain terms, all the AI giants already have an interest in developing technologies that comply with the rules and restrictions set by their creators. Market imperatives would constitute sufficient safeguards. Zuckerberg rejects the idea of slowing down and setting up a regulatory system. Alignment—that is, the ability of AI to align itself with human values and desires—would be a more than sufficient safeguard against all abuses. For Zuckerberg, “any lab that does not focus on alignment will fall behind”.

For the head of Meta, everything is fine

The creator of Facebook believes that AI giants do not need new laws to take precautions. They are plenty afraid of being dragged into court, which again constitutes a safeguard. To illustrate his point, he recalls that Meta chose to delay the release of Muse AI, its system capable of creating personalized agents, by several months, taking the time to better lock down the tool’s security.

Last month I wrote about how we can build a positive and safe future for everyone: https://t.co/1KtONlqBs5

Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens.

The…

— Mark Zuckerberg (@finkd) September 15, 2026

According to the billionaire, everything is therefore already in place for artificial intelligence to cause damage.Contrary to Altman and Amodei, he had estimated that summer that nothing should delay the arrival of new American AIs “not even by a month.”

Also read: OpenAI says it is ready to take a back seat on AI, but refuses to be the only one to do so

Mistral AI takes aim at its American rivals

In the wake of Mark Zuckerberg, Mistral AI, the French AI start-up, refuses to share the same view as Anthropic and OpenAI. In a statement sent to our colleagues at Le Figaro, the company first acknowledges the inherent dangers of artificial intelligence. It says that “AI models become powerful actors as soon as they are given room to maneuver in executing tasks and access to tools”. To deploy AI safely, it is necessary to exercise “genuine control” with “strict separation of access”

Above all, Mistral accuses, without naming anyone, the American AI giants of playing a double game by advocating new restrictions designed to drown out competition to their advantage. For the start-up, Anthropic and OpenAI use their repeated warnings to retain control of the AI market.

In any case, calls for regulation have not gained traction in Washington. Donald Trump firmly refuses to consider the idea of slowing down AI.The president fears that the slightest pause could allow China to get ahead of the United States. To drive the point home, Trump even called Jensen Huang, Nvidia’s CEO, in the middle of a conference.

👉🏻 Follow the latest tech news in real time: add 01net to your feeds on Google and subscribe to our WhatsApp channel.