AI assessment changes human behavior
Abstract
The shift from human to AI assessment raises an important question: do people behave differently when being assessed by AI? If so, this might have significant consequences for both people under assessment and the organizations conducting the assessment. Focusing on candidate selection decisions as a key assessment domain, the current research shows that AI assessment leads people to present themselves as more analytical because they believe that AI particularly values analytical characteristics. This behavioral shift could fundamentally alter who gets selected for positions, potentially undermining the validity of assessment processes. Overall, this work reveals how people change their behavior under AI assessment, offering insights for organizations and policymakers navigating the integration of AI in high-stakes decisions.
Study specs
Examined behaviors in candidate selection contexts to assess how people adapt their self-presentation under AI evaluation.
- Authors
- J Goergen,E de Bellis,AK Klesse
- Institution
- Cologne Business School,Maastricht University School of Business and Economics,Tilburg University,Copenhagen Business School
- Discipline
- Organizational Behavior,Psychology of AI
- Study Type
- Experimental Study
- Year
- 2025
- Human Data Platform
- Prolific
- Source
- View Source Google Scholar
Measured Outcomes
Changes in self-presentation and perceived traits emphasized during AI assessments compared to traditional evaluations.
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.