Do you want a say in what happens next?

PauseAI Belgium is a volunteer group asking for an international treaty to pause the training of the most powerful AI systems, until we know how to build them safely and keep them under democratic control.

It is coming fast

The big labs, researchers and independent forecasters expect AI as capable as the best humans at most tasks within a few years. Far more capable systems could follow quickly.

The problem

Nobody knows how to control it

Modern AI is grown, not built. In safety tests, current models cheat, lie, hide what they are doing and try to escape. Their makers cannot fully explain how they think.

Read the evidence

You can stop this. We know how.

An international pause on training the most powerful AI, verified the way nuclear treaties are. Humanity has held back technologies before, from CFCs to human cloning.

Our proposal

Upcoming events

No event is scheduled yet. Follow our calendar to hear about the next one.

PauseAI Belgium calendar PauseAI worldwide calendar

Watch and read

METR, August 2026 Investigation of the OpenAI / Hugging Face hacking incident An independent look at how the agents reasoned and worked together during the first known AI breach of this kind. Read Yoshua Bengio and 100+ experts, backed by 30+ countries, February 2026 International AI Safety Report 2026 The closest thing to a scientific consensus on what frontier AI can do and where it fails. Read Future of Life Institute, signed by Hinton, Bengio, Russell and 77,000+ others, October 2025 Statement on Superintelligence Calls for a prohibition on superintelligence until it is proven safe and controllable and has public support. You can sign it. Read Daniel Kokotajlo, Scott Alexander and others, April 2025 AI 2027 The most detailed scenario of the next few years, written by former OpenAI researchers. Read The AI 2027 team, 2026 AI 2040: Plan A What a good outcome would look like, and what the US, China and the rest of the world would have to do to get there. Read ING Belgium survey, September 2025 Seven in ten Belgians want strict AI rules 71% of Belgians favour strict AI rules even at the cost of some benefits, the highest among neighbouring countries. Read Eliezer Yudkowsky and Nate Soares, September 2025 If Anyone Builds It, Everyone Dies The bestselling case that building superhuman AI with today's methods ends badly. Read European Commission, July 2025 The EU General-Purpose AI Code of Practice The EU's own safety rules for the most advanced models, written a few streets from where we demonstrate. Read

In times of need, humanity comes together. We must only make sure it is not too late.

A PauseAI Belgium volunteer
Why we joined

Worried about AI? You can make a difference.

We are volunteers from all over Belgium, in Dutch, French and English. We demonstrate, hand out flyers, write to politicians and meet every week. No experience needed.

Join us