Illusions of truth—Experimental insights into human and algorithmic detections of fake online reviews

Abstract

The issue of fake online reviews is increasingly relevant due to the growing importance of online reviews to consumers and the growing frequency of deceptive corporate practices. It is, therefore, necessary to be able to detect fake online reviews. An experiment with 1041 respondents allowed us to create two pools of reviews (fake and truthful) and compare them for psycholinguistic deception cues. The resulting automated tool accounted for review valence and incentive and detected deceptive reviews with 81% accuracy. A follow-up experiment with 407 consumers showed that humans have only a 57% accuracy of detection, even when a deception mindset is activated with information on cues of fake online reviews. Therefore, micro-linguistic automated detection can be used to filter the content of reviewing websites to protect online users. Our independent analysis of reviewing websites confirms the presence of dubious content and, therefore, the need to introduce more sophisticated filtering approaches.