The ultimate bystander: A theoretical framework for trust-based AI intervention in workplace incivility
Itzkovich, Y., Aviv, Y., & Barzilai, O.
Human-Computer Interaction
AI Bystander Incivility Conceptual
TL;DR
If humans freeze in the face of workplace incivility, can a listening, intervening AI lower the threshold for help? A trust-based framework and research agenda.
Workplace incivility is pervasive, yet decades of bystander research show that human witnesses frequently fail to intervene, inhibited by diffusion of responsibility, fear of retaliation, social costs, and motivated empathy withdrawal. This paper develops a theoretical framework for AI as a trust-based "ultimate bystander" that can lower the threshold for help in the face of uncivil encounters. We review evidence on the psychology of bystander (in)action and on the organizational conditions that mute intervention, and argue that properly designed AI systems—combining passive observation, calibrated alerts, and simulation-based rehearsal—can address documented human barriers without supplanting human moral discretion. Drawing on theories of trust in automation, procedural justice, and human–AI teaming, we articulate the conditions under which employees, targets, and managers will accept an AI bystander as legitimate, and the conditions under which it will be perceived as surveillance and resisted. The framework specifies (a) the functional roles an AI bystander can play (detection, documentation, coaching, escalation), (b) the design principles needed to preserve autonomy and fairness, and (c) a research agenda with testable propositions linking trust, perceived justice, and intervention behavior. We close by considering ethical and organizational implications, including risks of misuse, bias, and over-reliance, and by proposing guardrails for responsible deployment in workplaces seeking to reduce incivility.
Cite
Itzkovich, Y., Aviv, Y., & Barzilai, O. (under review). The ultimate bystander: A theoretical framework for trust-based AI intervention in workplace incivility. Human-Computer Interaction.
Itzkovich, Y., et al. "The ultimate bystander: A theoretical framework for trust-based AI intervention in workplace incivility." Human-Computer Interaction, Under review.
Itzkovich, Y., Y. Aviv, and O. Barzilai. Under review. "The ultimate bystander: A theoretical framework for trust-based AI intervention in workplace incivility." Human-Computer Interaction.
@article{itzkovich2026,
title = {The ultimate bystander: A theoretical framework for trust-based AI intervention in workplace incivility},
author = {Itzkovich, Y. and Aviv, Y. and Barzilai, O.},
journal = {Human-Computer Interaction},
year = {2026},
note = {Under review}
}