Researchers poison their own data when stolen by an AI to ruin results

Researchers from China and Singapore developed AURA, a defense system that deliberately poisons proprietary knowledge graphs used in AI systems to prevent data theft. The method causes stolen knowledge graphs to produce hallucinations and wrong answers when used by thieves, while legitimate users with a secret key can still access correct information. Tests demonstrated 94% effectiveness in degrading stolen data utility, offering protection for valuable intellectual property in GraphRAG systems.