Watch Me Improve—Algorithm Aversion and Demonstrating the Ability to Learn

Abstract

Owing to advancements in artificial intelligence (AI) and specifically in machine learning, information technology (IT) systems can support humans in an increasing number of tasks. Yet, previous research indicates that people often prefer human support to support by an IT system, even if the latter provides superior performance – a phenomenon called algorithm aversion. A possible cause of algorithm aversion put forward in literature is that users lose trust in IT systems they become familiar with and perceive to err, for example, making forecasts that turn out to deviate from the actual value. Therefore, this paper evaluates the effectiveness of demonstrating an AI-based system’s ability to learn as a potential countermeasure against algorithm aversion in an incentive-compatible online experiment. The experiment reveals how the nature of an erring advisor (i.e., human vs. algorithmic), its familiarity to the user (i.e., unfamiliar vs. familiar), and its ability to learn (i.e., non-learning vs. learning) influence a decision maker’s reliance on the advisor’s judgement for an objective and non-personal decision task. The results reveal no difference in the reliance on unfamiliar human and algorithmic advisors, but differences in the reliance on familiar human and algorithmic advisors that err. Demonstrating an advisor’s ability to learn, however, offsets the effect of familiarity. Therefore, this study contributes to an enhanced understanding of algorithm aversion and is one of the first to examine how users perceive whether an IT system is able to learn. The findings provide theoretical and practical implications for the employment and design of AI-based systems.

Human vs. AI: Investigating Consumers’ Context-Dependent Purchase Intentions for Algorithm-Created Content

“…This attitude toward algorithms in a particular task can range from a strong aversion to an appreciation of algorithmic conduct. Previous research has, however, come to contrary conclusions about the circumstances in which algorithm aversion occurs [7,10].…”

“…The studies in these fields provide conflicting evidence. While some find a general aversion toward algorithms, others speak of an appreciation of algorithms [7,10]. There is already first evidence that, regardless of the content's actual author, the mere disclosure of algorithmic authorship leads to significant differences in consumers' perception of the content [1].…”

News from Artificial Intelligence is Believed Less

“…This research makes two theoretical contributions. First, it builds on and extends the literature on psychological responses to automated systems and on human-machine interaction (Dawes, 1979;Grove & Meehl, 1996;Meehl, 1954), indicating new -and perhaps surprisingdepths to this AI aversion given that past scholarship has shown cases of both AI appreciation (Berger et al, 2021;Castelo, Bos, & Lehman, 2019;Logg et al, 2019;Longoni & Cian, 2020) and aversion (e.g., Cadario, Longoni, & Morewedge 2021;Dietvorst & Bharti, 2020;Dietvorst, Simmons & Massey, 2015;Longoni, Bonezzi, & Morewedge, 2019). Though this literature spans over decades, it has mostly focused on outcome variables such as stated or reveal preference.…”

“…These qualities might be incorporated into how people judge content generated by AI-as being relayed impartially and dispassionately or, in other words, as more truthful. Furthermore, people appreciate algorithms more than humans for tasks that are impersonal (Berger et al 2021), require objectivity (Castelo, Bos, & Lehmann, 2019) or impartiality (Jago & Laurin 2021). As people want journalism to be impartial and neutral (American Press Institute, 2018), this AI appreciation account predicts that people would perceive news from AI as more accurate than news from human reporters, with higher trust ascribed to AI than human reporters.…”

News from Generative Artificial Intelligence is Believed Less

“…This research makes two theoretical contributions. First, it builds on and extends the literature on psychological responses to automated systems and on human-machine interaction (Dawes, 1979;Grove & Meehl, 1996;Meehl, 1954), indicating new -and perhaps surprisingdepths to this AI aversion given that past scholarship has shown cases of both AI appreciation (Berger et al, 2021;Castelo, Bos, & Lehman, 2019;Logg et al, 2019; and aversion (e.g., Cadario, Longoni, & Morewedge 2021;Dietvorst & Bharti, 2020;Dietvorst, Simmons & Massey, 2015;Longoni, Bonezzi, & Morewedge, 2019). Though this literature spans over decades, it has mostly focused on outcome variables such as stated or reveal preference.…”