For What It’s Worth: Humans Overwrite Their Economic Self-interest to Avoid Bargaining With AI Systems

Abstract

As algorithms are increasingly augmenting and substituting human decision-making, understanding how the introduction of computational agents changes the fundamentals of human behavior becomes vital. This pertains to not only users, but also those parties who face the consequences of an algorithmic decision. In a controlled experiment with 480 participants, we exploit an extended version of two-player ultimatum bargaining where responders choose to bargain with either another human, another human with an AI decision aid or an autonomous AI-system acting on behalf of a passive human proposer. Our results show strong responder preferences against the algorithm, as most responders opt for a human opponent and demand higher compensation to reach a contract with autonomous agents. To map these preferences to economic expectations, we elicit incentivized subject beliefs about their opponent's behavior. The majority of responders maximize their expected value when this is line with approaching the human proposer. In contrast, responders predicting income maximization for the autonomous AI-system overwhelmingly override economic self-interest to avoid the algorithm. CCS CONCEPTS• Human-centered computing → Empirical studies in HCI; User studies; Empirical studies in collaborative and social computing;• Applied computing → Economics; Psychology.

Humans Forgo Reward to Instill Fairness into AI

“…Moreover, individuals tend to experience less guilt when making unfair offers to AI than to another human (de Melo, Marsella, and Gratch 2016). Furthermore, individuals generally prefer interacting with other humans over AI and may change their responses to avoid AI interaction altogether (Erlei et al 2022). These findings suggest that if AI is trained using human interactions, it should account for existing human biases as well as behavioral changes driven by these interactions.…”

Gesticulate for Health’s Sake! Understanding the Use of Gestures as an Input Modality for Microtask Crowdsourcing

“…Artificial intelligence (AI) techniques and machine learning (ML) in particular, are drastically changing our lives through technological revolutions across several domains (Erlei et al 2020(Erlei et al , 2022. A primary stimulant that has led to these rapid advances in AI and ML in recent times, apart from the computational resources now available, is the design and development of crowdsourcing methods over the last decades to harness human intelligence at scale (Gray and Suri 2019).…”

Editorial: User Modeling and Recommendations

“…Frequently, these intelligent systems take advantage of psychological models to explain and predict user interactions with the systems (Tkalcic and Chen, 2015 ), influence user interaction through novel interfaces (Gupta et al, 2022 ), and allow for a deeper understanding of user behavior (Wölbitsch et al, 2019 ), including user trust in the systems (Erlei et al, 2020 ), and their reliance on such systems (Tolmeijer et al, 2021 ; Erlei et al, 2022 ), user preferences and needs (Wölbitsch et al, 2020 ; Najafian et al, 2021 ), which in turn also allow for more generalizable results. In complement, digital behavior in other systems has also been used to infer user characteristics.…”