Loading page…
Experiments linked the trust penalty to worse performance ratings and promotion assessments. Explicit encouragement reduced the penalty but did not eliminate it.
Listen to this story
In experiments reported in The Leadership Quarterly, supervisors were less willing to trust employees who disclosed using AI, with trust tied to evaluations of performance, contribution and promotion suitability. The researchers found that explicit company encouragement reduced—but did not erase—the penalty; having no AI policy produced a penalty similar to an outright ban. The results concern supervisors’ judgments in experimental scenarios, not real promotion decisions, and point to clear rules and manager training as ways to reduce incentives for employees to hide AI use.
Supervisors were more likely to trust employees who did the work themselves than employees who disclosed AI use: 68% versus 24%.
The experiments asked supervisors whether to verify employees’ reported work time, rather than measuring actual workplace promotions.
Lower trust was associated with poorer performance assessments, smaller perceived organizational contributions and lower promotion suitability.
An employer’s invitation to use AI may not protect employees from being judged for following it. A newly reported study in The Leadership Quarterly found that participants acting as supervisors trusted AI-using employees less—even when company policy encouraged the tools. Lower trust was also linked to worse evaluations and lower perceived suitability for promotion.
HealthDay detailed the findings on November 9 in an article published by U.S. News. It places the research in the journal’s December 2026 issue and cites a November 4 news release from the Hebrew University of Jerusalem. The experiments examined how disclosure of AI use changed supervisors’ willingness to trust an employee, rather than tracking actual promotions.
Participants supervised employees hired to create social media posts. They were told whether those employees had used AI, then faced a choice: accept the employees’ accounts of how much time they had worked, or request verification.
Futurism’s November 7 account says the supervisors were not basing that judgment on the quality of the work itself. It identifies the study’s lead researcher as Samuel Meyers, a Hebrew University doctoral student.
Reported chance of earning supervisors’ trust when deciding whether to verify work time.
Reported chance of earning supervisors’ trust when AI use was disclosed.
Additional experiments connected lower trust with poorer assessments of performance, smaller perceived contributions to the organization and lower perceived suitability for promotion.
The researchers also examined the rules surrounding AI use. Explicit company encouragement substantially reduced the loss of trust, but did not eliminate it. The penalty also appeared when AI was not explicitly prohibited. In their paper, the researchers said having no policy produced trust penalties similar to those seen when the technology was actively banned.
Senior researcher Yochanan Bigman, a Hebrew University senior lecturer, framed adoption as a psychological and managerial challenge as well as a technological one. An employee may produce better work more efficiently with AI and still lose a manager’s trust, he said.
The team’s proposed response combines clear AI policies with training managers not to penalize its use. Researchers warned that otherwise employees could keep using AI while concealing it—a practice they called shadow AI.
Loading discussion...
Join the conversation
Explain what would make that scrutiny justified or unfair.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.