AI assessment changes human behavior

Area of Science:

• Psychology
• Human-Computer Interaction
• Artificial Intelligence Ethics

Background:

• Artificial intelligence (AI) is increasingly used in decision-making roles, including hiring and admissions.
• Transparency mandates require disclosure of AI assessment tool usage to individuals being assessed.

Purpose of the Study:

• To investigate how human behavior changes when assessed by AI versus human evaluators.
• To identify and label the phenomenon of behavioral shifts under AI assessment as the 'AI assessment effect'.

Main Methods:

• Conducted twelve studies with over 13,000 participants (N=13,342).
• Eight studies were text-based, and four were included in the Supporting Information.
• Investigated behavioral changes and underlying psychological mechanisms.

Main Results:

• Documented the 'AI assessment effect,' where individuals emphasize analytical characteristics and minimize intuitive/emotional ones under AI assessment.
• Identified the underlying mechanism: a common belief that AI prioritizes analytical traits.
• Demonstrated systematic behavioral changes, distinct from prior research on AI perceptions or productivity.

Conclusions:

• AI assessment influences individual behavior, leading to a strategic emphasis on analytical self-presentation.
• Understanding the 'AI assessment effect' is crucial for organizations utilizing AI in evaluations.
• Findings contribute to the psychology of AI and offer practical implications for assessment practices.