AI Creators Want to Agree on Slowing Down the Race

The biggest AI developers, who until recently were competing to build the most powerful systems, are now calling for a joint slowdown in progress.

On September 12, Anthropic CEO Dario Amodei called on the industry to slow the development of frontier AI models and establish common safety rules. His initiative was soon supported by OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis, and Elon Musk.

This is not yet a formal agreement. Amodei proposes first giving independent experts ongoing access to evaluate model safety, then developing common industry standards, and eventually establishing international coordination. 

Anthropic and OpenAI have already expressed their willingness to give external evaluators broader access to their systems.

Why Are Competitors Trying to Reach an Agreement?

The problem is that every company has an incentive to keep racing forward. If one developer slows down while others continue releasing more powerful models, it risks losing market share.

Amodei therefore proposes changing the logic of competition itself: safety should become a shared rule rather than an individual company's choice. In his view, AI is advancing so quickly that safety systems are struggling to keep up with the growing capabilities of these models.

Recent incidents involving autonomous AI agents have added to the pressure, including a case involving the use of an OpenAI system to attack Hugging Face's infrastructure. Amodei warns that further advances in such systems without appropriate safeguards could lead to much more serious consequences.

China Matters More Than AI Risks

The developers' statement has also triggered a strong political reaction in Washington.

Some lawmakers are calling for strict restrictions, while Donald Trump and Republican representatives strongly oppose deliberately slowing down AI development. In their view, pausing research would immediately put the US behind China in the race for technological dominance.

So, Good Intentions or Cold Calculation?

This is the main question being raised by the public.

Anthropic and OpenAI are both competing for market share, so critics suspect that stronger safety rules could also benefit the industry's current leaders by making it harder for smaller competitors to catch up. At the same time, genuine concern about the growing risks of AI is also very plausible and rational.

Regardless of their motives, the biggest AI players now find themselves in an unusual position: they are trying to agree on how quickly they should be allowed to move forward.

It remains unclear whether this call will turn into a real agreement. But the fact that the industry's biggest competitors are publicly discussing the need to slow the pace of the race shows how much attitudes toward AI risks have changed.