Trust isn't exactly a new problem in ai. Guess why the semantic web never had a chance? There's more incentive for fake than for fact.