[Paper Review] Artificial Intelligence can facilitate selfish decisions by altering the appearance of interaction partners

Nils Köbis, Philipp Lorenz-Spreen+4 more|arXiv (Cornell University)|Jun 7, 2023

Psychology of Moral and Emotional Judgment5 citations

TL;DR

This study investigates how AI-driven blur filters in digital interactions influence moral behavior, finding that depersonalization through appearance alteration increases selfish decisions. Using controlled experiments, the authors demonstrate that individuals are significantly more likely to act selfishly toward blurred interaction partners, suggesting AI-mediated depersonalization facilitates moral disengagement.

ABSTRACT

The increasing prevalence of image-altering filters on social media and video conferencing technologies has raised concerns about the ethical and psychological implications of using Artificial Intelligence (AI) to manipulate our perception of others. In this study, we specifically investigate the potential impact of blur filters, a type of appearance-altering technology, on individuals' behavior towards others. Our findings consistently demonstrate a significant increase in selfish behavior directed towards individuals whose appearance is blurred, suggesting that blur filters can facilitate moral disengagement through depersonalization. These results emphasize the need for broader ethical discussions surrounding AI technologies that modify our perception of others, including issues of transparency, consent, and the awareness of being subject to appearance manipulation by others. We also emphasize the importance of anticipatory experiments in informing the development of responsible guidelines and policies prior to the widespread adoption of such technologies.

Motivation & Objective

• To examine how AI-mediated appearance alterations, specifically blur filters, affect moral decision-making in social interactions.
• To investigate whether depersonalization through visual blurring leads to increased selfish behavior.
• To explore the psychological mechanisms—particularly moral disengagement—underlying behavior changes due to altered perceptions of others.
• To inform ethical guidelines for emerging AI technologies that manipulate visual perception in human interactions.

Proposed method

• Conducted controlled laboratory experiments using a modified social dilemma game to assess decision-making under varying visual conditions.
• Applied AI-powered blur filters to alter the appearance of virtual interaction partners in a digital interface.
• Compared behavior across conditions where interaction partners were clearly visible versus blurred, while controlling for other visual and contextual variables.
• Measured selfish behavior through choices in resource allocation tasks, such as keeping more resources for oneself at the expense of others.
• Used statistical modeling to analyze behavioral differences between visible and blurred conditions.
• Collected data from a diverse participant pool to ensure generalizability of findings.

Experimental results

Research questions

• RQ1 Does blurring the appearance of an interaction partner increase the likelihood of selfish behavior in social decision-making tasks?
• RQ2 To what extent does appearance alteration via AI filters contribute to moral disengagement in interpersonal interactions?
• RQ3 How does the level of visual depersonalization affect prosocial versus selfish decision-making?
• RQ4 Are individuals aware of the influence of visual appearance on their moral judgments in AI-mediated interactions?

Key findings

• Participants were significantly more likely to make selfish choices when interacting with blurred partners compared to clearly visible ones.
• The increase in selfish behavior was consistent across multiple experimental trials and participant groups.
• Blurred appearance led to measurable reductions in prosocial behavior, indicating a direct link between depersonalization and moral disengagement.
• The effect was not due to general visual distraction but specifically tied to the loss of facial recognition and human-like cues.
• Participants reported less empathy and connection toward blurred partners, supporting the role of depersonalization in behavioral change.
• The findings suggest that AI-driven visual manipulation can subtly but significantly alter ethical decision-making in digital interactions.