Hi HR Division Colleagues,
I wanted to share some recent "in the wild" field data that has significant implications for those of us researching recruitment, selection, and the future of work.
My colleagues and I just completed a multi-wave study of over 9,000 real-world job applicants. We found that self-reported GenAI use in unproctored assessments has skyrocketed from 3% to over 23.3% in just twelve months.
We also ran a Randomized Controlled Trial (RCT) to see if we could mitigate this "arms race" without resorting to expensive (and often flawed) AI-detection software. Our findings suggest that a simple behavioral nudge-a consequences-based warning screen-cut AI-assisted faking by more than half, reducing incidence to 10.6%.
Interestingly, the data also shows that these warnings don't hurt the candidate experience; in fact, they may improve perceptions of procedural justice by providing clear rules of engagement.
The full study was recently published in the International Journal of Selection and Assessment (Open Access): <response-element class="" ng-version="0.0.0-PLACEHOLDER"></response-element>https://doi.org/10.1111/ijsa.70056<response-element class="" ng-version="0.0.0-PLACEHOLDER"></response-element>
I'd love to hear from others in the division:
-
How are you or your partner organizations adjusting selection hurdles in light of this 20%+ surge?
-
Are we moving toward a "proctored-only" world, or can behavioral nudges save the unproctored assessment?
Looking forward to the conversation.
Chet Robie, Ph.D. SIOP Fellow | Wilfrid Laurier University
------------------------------
Chet Robie, PhD
Professor
Wilfrid Laurier University
Lazaridis School of Business & Economics
Waterloo ON
crobie@wlu.ca------------------------------