What happened
A study led by researchers at the Hebrew University of Jerusalem and WHU—Otto Beisheim School of Management found that employees who use AI tools like ChatGPT face a significant 'trust penalty' from their supervisors. Across six experiments, supervisors were found to be less likely to trust work produced with AI assistance compared to work completed independently, even when the quality of the output was not the primary factor in the evaluation.
The research, published in The Leadership Quarterly, involved six separate experiments examining how the use of large language models (LLMs) influences supervisor perceptions. In one key experiment, supervisors were asked to evaluate reports produced either independently or with AI assistance. The results showed that employees who worked independently had a 68% chance of earning supervisor trust, while those using AI had only a 24% chance.
The study explicitly noted that supervisors were not evaluating the quality of the finished work, but rather the process of its creation. The researchers found that while AI-assisted work was viewed with less trust than independent work, it was still perceived more favorably than plagiarism.
The negative trust impact was observed across various contexts, including a study involving academic supervisors evaluating graduate student work. The researchers found that the more extensively students used AI, the less their supervisors tended to trust them.
Why it matters
This research highlights a critical psychological and managerial barrier to AI adoption in the workplace. While organizations often push for AI-driven productivity, the study suggests that the mere presence of AI in a workflow can negatively impact an employee's professional standing, including performance evaluations and promotion prospects. This creates a 'paradox of the AI revolution' where employees are incentivized to use tools that may simultaneously damage their perceived reliability. The findings suggest that without clear, supportive organizational policies, the use of AI could lead to 'shadow AI,' where employees hide their tool usage to avoid professional repercussions.
The study suggests that the integration of AI into the workplace is as much a psychological challenge as a technical one. Lower trust levels were directly associated with poorer performance evaluations, reduced perceptions of an employee's contribution, and lower perceived suitability for promotion.
The researchers warn that this tension creates a risk of 'shadow AI,' where employees conceal their use of technology to avoid negative judgments. This behavior could hinder organizational transparency and the effective management of AI-driven workflows.
The findings indicate that neither the importance of the task nor the supervisor's own experience with AI consistently eliminated the trust penalty, suggesting that the against AI-assisted work is deeply rooted in current managerial perceptions.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
What to watch next
The study identifies a potential mitigation strategy: explicit organizational encouragement. When companies formally endorse and encourage the use of AI, the negative impact on supervisory trust was found to be no longer statistically significant. Future workplace dynamics will likely depend on whether organizations can successfully transition from ambiguous or prohibitive stances to clear, supportive policies that normalize AI as a standard professional tool rather than a shortcut that undermines individual competence.
The study suggests that the most effective way to mitigate the trust penalty is for organizations to provide clear, explicit policies that encourage the use of AI. When such policies are in place, the negative effect on trust was found to be no longer statistically significant.
Organizations may need to implement specific managerial training to help supervisors distinguish between the appropriate use of AI as a productivity tool and concerns regarding employee competence.
The researchers emphasize that the benefits of AI adoption in the workplace may remain unrealized if employees fear that using these tools will negatively impact their career trajectory and professional reputation.