What is interesting to HN is that the police may have used AI summaries to determine the risk, and the AI/Search gave them the wrong information. Essentially they relied on Google or copilot (it is not clear which) to inform them of a riot at a match that didn't even take place. Police intelligence relying on public search and AI that is hallucinating is probably as common as it is worrying.