What happened
Researchers built a system that lets AI agents automatically filter what information they receive based on legal rules about data use — making it harder (not easier) for agents to access content they're not supposed to see. In practice, this means when multiple AI systems share a knowledge base, each one only gets notified about information it's legally allowed to use, preventing accidental violations of data regulations like the EU's AI Act.
Why it matters
As AI systems become more interconnected and share knowledge bases, they need built-in guardrails that enforce legal rules about who can use what data — this is the first concrete implementation showing that semantic filtering (matching relevant information) and regulatory compliance can work together in real time.