I like the idea of white hat AI detectives, but how that’s deployed, monitored, and controlled is crucial. There needs to be robust checking of the “checkers”, otherwise we end up in perilous situations like the NSA spying on millions of Americans without their knowledge.