The illusion of insight: detailed warnings reduce but do not prevent false "Aha!" moments

False "Aha!" moments can be elicited experimentally using the False Insight Anagram Task (FIAT), which combines semantic priming and visual similarity manipulations to lead participants into having "Aha!" moments for incorrect anagram solutions. In a preregistered experiment (N = 255), we tested whether warning participants and explaining to them exactly how they were being deceived, would reduce their susceptibility to false insights. We found that simple warnings did not reduce the incidence of false insights. On the other hand, participants who were given a detailed explanation of the methods used to deceive them experienced a small reduction in false insights compared to participants given no warning at all. Our findings suggest that the FIAT elicits a robust false insight effect that is hard to overcome, demonstrating the persuasive nature of false insights when the conditions are ripe for them.