Restricting the AI race, as was once done with nuclear weapons. Is that possible?

The Dutch government is pioneering an AI monitoring system to oversee data centers, aiming to enhance international AI safety discussions. This initiative, led by State Secretary Willemijn Aerdts, is seen as a significant step towards establishing trust among nations regarding AI development. The prototype technology can detect suspicious activities in data centers, ensuring compliance with potential international agreements on AI. Experts draw parallels with nuclear non-proliferation efforts, emphasizing the need for transparency and verification in AI advancements. The demonstration marks a first for government hardware verification in AI, highlighting the urgency of addressing AI-related risks.