Man vs. Machine: The Substitutability of AI and Expert Evaluations of Initial Coin Offerings (ICOs)
by
Lauren Rhue
Assistant Professor of Information Systems
Department of Decision, Operations and Information Technologies
Robert H. Smith School of Business
University of Maryland
Friday, Sep 24
9 – 10 am | Zoom
Abstact:
Initial coin offerings (ICOs) were heralded as a popular method for emerging blockchain and technology ventures to raise capital for their businesses; however, several high-profile ICO scams generated concerns about ICO legitimacy. We examine an ICO-rating platform that provides two evaluation sources, artificial intelligence (AI) and experts, as well as qualitative and quantitative expert evaluations to evaluate ICOs. This study compares the informativeness of the information sources and information types to understand how AI ratings for uncertain quality items, like ICOs, compares to the expert ratings. Using dual-processing theory and cognitive biases, we posit substitutability for quantitative evaluations but complementarity between quantitative and qualitative evaluations. Using nearly 5,000 ICOs and more than 14,000 expert evaluations, we find that 1) experts’ decisions on which ICOs to evaluate contain relevant information, 2) experts and AI quantitative evaluation are substitutes, and 3) quantitative evaluations complement qualitative evaluations. Our paper makes several contributions to the information systems literature related to the substitutability of automation systems for online human reviews, the different processing pathways for qualitative and quantitative evaluations, and the unexpected benefit of cognitive biases.