Over the past decade, artificial intelligence (AI) has seen tremendous progress. This has allowed organizations to automate decision-making processes traditionally handled by humans. However, many people react negatively to algorithmic decision-making, particularly when these decisions affect them at work. This resistance is especially evident in personnel selection, where both practitioners and applicants tend to view automated decision-making unfavorably. This article (Wesche et al., 2024) seeks to address these negative perceptions.
PERCEPTIONS OF AI DECISION-MAKING
In two studies, the researchers examined how people respond to algorithmic versus human decision-making. Study 1 involved a simulated selection scenario. Results showed that algorithmic decision-making negatively impacted participants’ trust and acceptance of decisions. However, when researchers provided participants with a clear explanation about how the algorithmic tool works, ratings of transparency, trust, and acceptance improved.
Study 2 involved participants who were competing for incentives, which made it an actual selection scenario. Here, the results confirmed that people generally prefer human decision-making over algorithmic approaches. However, providing participants with explanations did not alleviate the negativity associated with the algorithms.
Overall, this study suggests that people generally have negative reactions to algorithmic decision-making. Providing explanations about how the algorithm works may alleviate concerns in some situations; however, according to the authors, this finding is not conclusive and may require further research.
ORGANIZATIONAL TAKEAWAYS
Based on these insights, the authors offer the following recommendations:
- Organizations should carefully decide which tasks to assign to algorithms and which to keep under human control, as trust and acceptance often depend more on the perceptions of automation, rather than its actual performance.
- Clear communication about decision-making processes might improve trust and acceptance. Organizations could tailor information based on stakeholder groups and whether decisions are automated or made by people.
Wesche, J. S., Hennig, F., Kollhed, C. S., Quade, J., Kluge, S., & Sonderegger, A. (2022). People’s reactions to decisions by human vs. Algorithmic decision-makers: The role of explanations and type of selection tests. European Journal of Work and Organizational Psychology. Advance online publication.
Image credit: Unsplash+