Biased or not biased? Exploring consumer's performance in detecting biased recommendation agents

Recommendation agents (RAs) provide consumers with personalized product recommendations based on their preferences, helping them to purchase online. An increasing number of merchants use biased RAs on their online store providing recommendations not solely made to match consumers' preferences but biased towards their strategic goals. However, the literature does not investigate further whether consumers are aware of this phenomenon. Based on agency theory and signal detection theory, this paper conducts four experiments to investigate the effects of the level of bias intensity in RA's recommendations, perceived personalization, perceived usefulness, and consumer's experience with the RA on consumer's ability to detect biased RAs. Results are not yet available, but preliminary research shows results in line with hypotheses. By providing a better understanding of consumers' vulnerability to the online merchants’ business practices and the mechanisms underlying bias detection performance, the results offer a valuable contribution to research and practice.