The European Union began enforcement this week of what it bills as the world’s first comprehensive artificial intelligence law. It’s already outdated.
The bureaucrats in Brussels are struggling to keep up with the breakneck pace of technological progress. In 2024, when this passed, the law decreed that an AI model is presumed to pose a “systemic risk” if it’s trained using more than 10 to the power of 25 mathematical operations — a snapshot of where the frontier stood during negotiations. Models are now far more capable.