PauseAI demo ASML

​The Hugging Face attacks showed the world that AI is getting more and more powerful, while still being misaligned. Anything with an internet connection, whether this is infrastructure, hospitals, banks, governments, etc., is now at risk from rogue AI agents. And when recursive self-improvement (RSI) kicks in, which may be soon, we could end up with giantly more powerful rogue AIs.

Experts from the AI labs that understand the technology best, as well as Nobel-prize winning scientists, say there are very significant chances that humanity will lose control over AI, which may result in human extinction. Politicians ranging from Bernie Sanders to Barack Obama to Ursula von der Leyen, as well as authoritative voices from the Pope to the UN chief to Bill Gates, are all saying that we must take these warnings seriously. Meanwhile, the Trump administration refuses to do anything but self-regulation.

ASML is producing the technology that makes modern AI possible. Without their lithography machines, none of these models can be trained. ASML can therefore ask downstream chip manufacturers to enforce safety regulation upon their customers, making sure AI can only be developed in a safe and responsible manner.

At PauseAI, we are calling upon ASML's leadership and workers to:

- Recognize the existential risks that come with developing frontier AI.
- Recognize the unique responsibility ASML has to make sure these risks get managed globally.
- Ask its customers to enforce downstream AI safety policies, ruling out the irresponsible practices we see today.

Join us for a friendly, strictly legal and strictly nonviolent demonstration in Veldhoven to make our voices heard!