Use of artificial intelligence and concerns about that use were common themes in symposia at this year’s meeting of the Society for Industrial and Organizational Psychology. In the exhibit hall vendors were demonstrating new AI applications, including use in personnel selection. I enjoy walking through poster sessions, seeing what new research is underway, talking with researchers. There was one poster on AI that I found particularly interesting.
Manuel Gonzalez (doctoral candidate, Baruch College & The Graduate Center, CUNY) and his colleagues from Aon Assessment Solutions tried to assess job applicants’ possible attitudes toward use of AI in hiring decisions. They had nearly 200 MTurk workers respond to hypothetical situations and a variety of self-report attitudinal measures. Decision-makers were human or AI. Here were some of their findings:
- “People generally react unfavorably toward AI and the organizations that use them for selection purposes”
- People were less likely to trust organizations that used AI in decision-making.
- People were less likely to promote organizations that used AI in decision-making except people who were very familiar with AI.
- Where AI made the decisions, people had less trust, saw as potentially less fair, and had concerns about privacy and communications.
- Emphasize interpersonal interactions in the selection process when using AI.
- Explain how AI works and how it is used.
1. Our attitudes toward AI are evolving. Examining attitudes toward AI across many areas of application is smart. Gonzalez and colleagues are to be commended for their contribution. This should be a growing area for research.
2. It is a brave new world. AI can increase efficiencies and help make more effective decisions. Should AI be the deciding factor in personnel decisions? This question echoes Paul Meehl (1954) on the value of statistical prediction versus clinical prediction. Statistical will usually have superior outcomes. However Meehl made a caveat. The statistical prediction must be properly used. I think that is the issue with AI software. The devil is in the details. For example, is AI eliminating people from hiring consideration who might otherwise be good employees? Overheard in one session at the conference: “Perhaps we do not have a shortage of job applicants; perhaps AI is eliminated some good applicants.” Managers and leaders should be familiar with how their organizations use AI to advance objectives.
3. So how do the faculty who teach AI make decisions in hiring new faculty? Do they use AI? My guess is they do it like most faculty usually do -- going through a stack of applicant folders, making decisions to narrow the field, thus producing a smaller pool. Then repeating the process until two or three candidates are invited to campus for job interviews and teaching demonstrations. Then taking a vote.
4. Hiring new employees is one of the great opportunities to build and grow an organization. AI can make the process more efficient. Each organization, each leader, must decide how AI can help.
Gonzalez, M. F., Martin, N. R., Justenhoven, R., & Preuss, A. (2019, April). Rage against the machine: Reactions to artificial intelligence in selection systems. Poster session presented at the Annual Conference of the Society for Industrial and Organizational Psychology, National Harbor, MD.
Meehl, P. E. (1954). Clinical versus statistical prediction: A theoretical analysis and a review of the evidence. Minneapolis, MN, US: University of Minnesota Press.
Image by geralt. Obtained from https://pixabay.com/illustrations/artificial-intelligence-brain-think-3382507/
Free to use.
© John Ballard, PhD, 2019. All rights reserved.
Decoding the Workplace “deals with principles and practices that are timeless . . . Is this a must-have for managers and would-be managers? Yes.” Academy of Management Learning & Education, June, 2018.
New: Now available in paperback.