When and why do minorities prefer advice from AI versus human mentors in workplace settings?
Research Interests
The riseof big data has increased both the availability and utility of a new source ofadvice: algorithms. Algorithms are defined as “computational procedures drawingon some type of digital data that provide some kind of quantitative outputthrough a software program” (Christin, 2017, p. 2).
Algorithms,sometimes called Artificial Intelligence, are processes or operations given toa computer to compile lots of data into easy-to-understand information forhuman users. Some research has demonstrated that people distrust of algorithmicoutput, sometimes referred to as algorithm aversion (Dietvorst, Simmons,& Massey, 2015). Other research has also argued that people prefer advicefrom algorithms to advice from people, an effect referred to as algorithmappreciation (Logg, Minson, & Moore, 2019).
In thisproject, we aim to reconcile these two findings by examining when and whyminorities might prefer advice from algorithms versus other people in workplacesettings. Specifically, we will investigate whether the user’s own genderand/or ethnicity would influence whether they perceive algorithms to be helpfulor not. We will also investigate whether the type of advice (i.e., instrumentaladvice over affective advice) would alter people’s preferences. Overall, wehypothesize that minority groups (e.g., women and racial/ethnic minorities) mayprefer advice from algorithms because it mitigates the fear of confirmingstereotypes.
Desired outcomes
By leveraging our collective expertisein social psychology, public policy and organizational behavior, we seek tobetter understand how emerging technologies such as algorithms can influenceworkplace outcomes for minorities. Through this inquiry, we will developinsights that can enable us to advocate future technology-based interventionsfor diversity and inclusion efforts that are critical in organizations and oursociety today. We plan to build a research program in this area by combininginsights from social psychology, public policy, and organizational behavior.Because this project has implications across multiple fields, we consider it tobe truly interdisciplinary in its approach and contribution. We believe this projectwill also allow us to secure external funding (e.g., NSF, NIH) to furtherpursue a comprehensive research program on this topics and result in academicand popular press publications. Thefindings from this project will be especially relevant to various researchstreams given the increased interest in algorithms. Finally, we also see thiswork as contributing to our understanding of diversity and inclusion efforts.
Student engagement
Studentsare an integral part of this project. We expect to allocate 50% of the fundingto support Postdoctoral Researcher at Batten where she can hire several undergraduate andgraduate research assistants who are interested in social psychology,organizational behavior, judgment and decision making, and public policy. RAs willbe exposed to all levels of projects from idea development to deliverables(e.g., paper and presentation), in which they will gain experience in socialscience research. We expect to allocate the remaining 50% of the funding tohuman participant incentives and study materials. There is a potential to hirean external programmer to create real-time algorithms for experimentalpurposes, which will be covered by the funding.