Is the race to build AI moving too fast?

By Lauren Wilkin

A petition to slow down AI development and put humans before technological advancements has reached almost 1 million signatures.

The petition is gaining even more recognition since a former researcher for Anthropic announced he believed there was a 10% chance AI will wipe out the human race within 10 years.

The organization, Stop the Race to Replace Humans, was established this year, and is calling on a "Pro-Human AI Declaration" to be signed.

The declaration sets out that artificial intelligence should serve humanity, not the reverse.

The petition states: "Every powerful technology in history has had an off-switch. AI companies are racing to build systems that might not.

"Nuclear plants have shutdown protocols. Airplanes have manual overrides. Even your smoke detector has a test button.

"But some of the most powerful AI systems being built right now are explicitly designed to resist being stopped, to improve themselves without human sign-off, and to operate faster than any human can review.

"The more people who sign, the harder it becomes for a few billionaires to decide humanity's future without us."

The petition now stands at 925,471 signatures, and it only needs 74,529 more to reach 1 million.

More from this section

+7

How melting Arctic glaciers is creating a climate ‘doom loop’

Learning without a brain – how bacteria store memories and remember the past like artificial neural networks

Pennsylvania is suing TikTok over ‘addictive’ features – a policy expert explains why it’s messier than that sounds

It's currently being widely circulated after Evan Hubinger — a researcher for Anthropic — replied to a post on X and said: "Jacob (Coxon) is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to."

And today, Anthropic has released its threat report of 2026, that analyzes the threat posed by its technology and how it is being used.

The report revealed a number of shocking discoveries.

Analysis of use cases revealed that a Russian-linked espionage actor used Claude to automate elements of the attack.

Another example of misuse is in Malaysia's political infrastructure. AI was used to analyze census and electoral information, create and operate over 1,000 fake X accounts, and then generate political content and allegations that circulated widely.

Anthropic has also identified the risk of the development of biological weapons. Multiple cases of deep research into viruses were uncovered — and even though they cannot prove it is specifically with intent to develop weapons — one case identified was conducted by a military institute.

This military institute — who have not been named — made a request to AI concerning "gain-of-function" work on chikungunya virus.

This generally refers to deliberately changing a virus to give it a new or enhanced characteristics.

Chikungunya virus is a mosquito-born virus, and the government website warns it can cause muscle pain, headache, nausea and fatigue.

The report comprises 154 pages of similar cases spanning from December 2025 to September 2026 that pose a threat to humanity.