Lie detection algorithms disrupt the social dynamics of accusation behavior

“The aggregate results across treatments suggest the following conclusions: (1) in the absence of a lie-detection algorithm, people are reluctant to make accusations; (2) when the lie-detection algorithm is available, a minority of people want to obtain its prediction; (3) the minority that does almost always follows the algorithmic prediction, independent of whether the algorithm flags a statement as truth or lie; (4) individuals who actively request algorithmic predictions are not inherently more prone to make accusations, but follow accusation suggestions more than those who receive such predictions without actively seeking them; (5) those who would not actively request the algorithmic prediction do not change their behavior even when they passively receive one; (6) beliefs about the relative performance of the lie-detection algorithm correlate with adoption rates.”

Von Schenk, A. et al. (2024) ‘Lie detection algorithms disrupt the social dynamics of accusation behavior,’ iScience, p. 110201. https://doi.org/10.1016/j.isci.2024.110201.

#OpenAccess #OA #Research #Article #DOI #Computing #SocialScience #Social #MachineLearning #Algorithm #Lying #Academia #Academic #Academics @socialscience