Employees who leverage artificial intelligence (AI) tools such as ChatGPT in their professional roles may encounter a significant deficit in trustworthiness as perceived by their supervisors, a factor that could critically impede performance evaluations and career advancement. This stark finding emerges from a comprehensive six-study research paper recently published in The Leadership Quarterly, shedding new light on the nuanced dynamics of human-AI collaboration in organizational settings. The investigation, which combined five incentivized economic experiments with a targeted field study involving academic supervisors, meticulously explored the extent to which the use of large language models (LLMs) by employees influences managerial trust.
The Trust Deficit: Quantifying Managerial Perceptions
One of the pivotal experiments detailed in the research paper presented supervisors with a scenario where they had to decide whether to accept an employee’s self-reported time for task completion or independently verify the information. The results painted a clear picture: employees who completed their tasks autonomously garnered supervisors’ trust in 68% of instances. In sharp contrast, when employees utilized AI-generated output for the same task, the likelihood of receiving supervisory trust plummeted dramatically to just 24%. This significant drop underscores a deep-seated apprehension among managers regarding AI-assisted work.
Crucially, the study also provided a comparative perspective, indicating that while AI-assisted work was viewed with less trust than independently produced work, it was still perceived more favorably than work explicitly copied from another person. This nuance suggests a hierarchy of mistrust, with direct plagiarism being the most egregious, followed by AI-generated content, and then the gold standard of independent human effort. The researchers were careful to highlight that their experiment specifically focused on supervisors’ willingness to trust employees’ accounts, rather than evaluating the quality of the final output. This distinction is vital, as it implies that the mere knowledge of AI involvement, irrespective of the work’s actual standard, was a primary driver in shaping perceptions of employee trustworthiness. The implications of this finding extend far beyond immediate trust, permeating various aspects of professional evaluation.
Unpacking the Consequences: Beyond Immediate Trust
The cascading effects of lower supervisory trust, as revealed across the suite of studies, are profound and multifaceted. Reduced trust was consistently associated with poorer perceived performance by supervisors. This suggests that even if the quality of AI-assisted work is high, the perception of its origin can bias managerial judgment against the employee. Furthermore, lower trust correlated with diminished assessments of organizational citizenship behaviour—a measure of an employee’s voluntary commitment to the organization beyond their formal job requirements. Such behaviours, crucial for a healthy workplace culture, appear to be undermined when trust erodes. Perhaps most critically for career trajectories, the research also found a direct link between reduced supervisory trust and lower perceptions of employees’ promotability, effectively stalling potential advancement opportunities.
The Crucial Role of Organizational Policy
Amidst these concerning findings, the research identified a significant mitigating factor: organizational policies. The studies revealed that when companies explicitly encouraged employees to integrate and utilize LLMs in their work, the negative impact on supervisory trust was no longer statistically significant. This suggests that a clear, positive stance from leadership can effectively neutralize the inherent distrust managers might otherwise feel. However, the "trust penalty" remained firmly in place when organizations either explicitly prohibited AI use or, more commonly, maintained an ambiguous or absent policy regarding its adoption. This highlights a critical need for proactive policy development rather than reactive bans or passive non-intervention.
Interestingly, the researchers also investigated whether factors such as the importance of the task at hand or the supervisors’ own prior experience with AI tools influenced their reactions. Neither of these variables consistently managed to reduce the observed negative effect on trust, reinforcing the pervasive nature of this managerial apprehension. A complementary field study, involving academic supervisors evaluating the work of graduate students, corroborated these findings, demonstrating that broader student use of LLMs corresponded directly with lower reported trust from their supervisors.
The Rise of Generative AI: A Brief Chronology of Workplace Integration
The rapid ascent of generative AI tools like ChatGPT since its public launch in November 2022 has marked a pivotal moment in technological history. Initially greeted with a mix of awe and trepidation, these powerful language models quickly permeated various sectors, promising unprecedented gains in productivity, creativity, and efficiency. Early adopters, both individuals and organizations, enthusiastically experimented with AI for tasks ranging from drafting emails and reports to generating code, brainstorming ideas, and analyzing data.
However, this initial enthusiasm was soon tempered by a growing awareness of potential pitfalls. Concerns around data privacy, intellectual property rights, the potential for misinformation, and ethical implications began to surface. Organizations found themselves in a challenging position, grappling with how to integrate these transformative tools safely and effectively, often in the absence of established best practices or regulatory frameworks. Many adopted a wait-and-see approach, while others implemented immediate, often temporary, bans on AI tools to prevent data leaks or compliance breaches. This period of rapid innovation juxtaposed with policy vacuums set the stage for the trust challenges now being uncovered. The lack of clear guidance from leadership often left employees to navigate the nascent AI landscape on their own, leading to inconsistent adoption patterns and varied managerial reactions.
Navigating the Policy Labyrinth: Organizational Responses to AI
The organizational world’s response to the AI revolution has been a dynamic and evolving process. In the immediate aftermath of ChatGPT’s widespread availability, many companies, particularly those in sensitive sectors like finance, legal, and healthcare, opted for outright bans on generative AI tools. This was primarily driven by fears of data breaches, intellectual property leakage, and regulatory non-compliance. These early, restrictive measures, while understandable from a risk management perspective, often overlooked the potential benefits of AI and could inadvertently stifle innovation within their workforces.
As understanding of AI’s capabilities and limitations grew, a shift began to occur. Forward-thinking organizations recognized that outright prohibition was unsustainable and could put them at a disadvantage. The focus began to move towards cautious exploration, piloting AI tools within controlled environments, and developing internal guidelines. This gradual transition reflected a growing realization that AI was not a fad but a fundamental shift in how work would be performed. The Leadership Quarterly study’s finding—that explicit organizational encouragement of LLM use negated the trust penalty—underscores the critical importance of this policy evolution. It suggests that when leadership endorses AI, it signals legitimacy and reduces the perceived risk or impropriety associated with its use, thereby fostering an environment where trust can be maintained. This places a significant burden on HR departments and senior management to craft nuanced policies that balance innovation, productivity, risk mitigation, and, crucially, the cultivation of trust.
The Shadow AI Phenomenon: A Crisis of Disclosure
The findings of this research also illuminate a growing concern known as "shadow AI." This phenomenon describes instances where employees utilize AI tools in their work without formally disclosing their usage to their supervisors or the organization. The study’s results offer a compelling explanation for its prevalence: employees may feel compelled to conceal their AI use if they believe that admitting to it could damage their credibility, lead to lower performance assessments, or hinder their career prospects. This creates a vicious cycle where a lack of trust fuels secrecy, which in turn can further erode trust if discovered.
Shadow AI carries significant risks for organizations. It can lead to the unsanctioned processing of sensitive company data by third-party AI models, creating cybersecurity vulnerabilities and compliance headaches. It can also result in inconsistencies in work quality, attribution issues, and a lack of transparency that hinders effective oversight and collaboration. From an employee perspective, the constant need to hide AI use can be a source of stress and anxiety, potentially impacting their well-being and overall job satisfaction. The existence of shadow AI highlights a fundamental disconnect between employee desires for efficiency and organizational readiness to support AI integration transparently. Addressing this requires not just policy, but a culture of open communication and psychological safety.
Broader Implications for the Future of Work and Leadership
The implications of this research extend far into the broader landscape of the future of work and leadership. As AI tools become increasingly sophisticated and ubiquitous, they are fundamentally reshaping job roles, skill requirements, and organizational structures. This study emphasizes that beyond the technical integration of AI, the human element—specifically trust—remains paramount.
For organizations, the findings suggest that merely introducing AI tools without clear guidelines is insufficient. A strategic approach is required, focusing on transparent communication with managers and employees about acceptable use, ethical considerations, and the role of AI in enhancing human capabilities. This will involve:
- Redefining Productivity and Accountability: In an AI-augmented workplace, what constitutes "productivity" shifts. Managers need to understand how to evaluate work where AI has played a significant role and how to attribute accountability fairly.
- Evolving Role of Human Supervision: Managers’ roles will transform from direct oversight of manual tasks to guiding and coaching employees on effective AI collaboration, ensuring ethical use, and fostering critical thinking skills to validate AI outputs.
- Adapting Performance Management Frameworks: Existing performance review systems may need significant overhaul to accommodate AI assistance. Metrics and evaluation criteria must be updated to reflect the new ways work is performed, recognizing AI as a tool rather than a substitute for human effort.
- Cultivating an AI-Positive, Trust-Centric Culture: Leadership must actively foster an environment where AI is seen as an augmentation tool, not a threat to human value or a reason for suspicion. This requires open dialogue, education, and consistent messaging from the top down.
Expert Perspectives and Recommendations
The researchers behind this seminal paper underscore the urgency for organizations to address these trust dynamics proactively. Their recommendations are clear and actionable:
- Develop Clear Guidelines: Organizations must establish explicit, unambiguous policies regarding the acceptable use of LLMs and other AI tools. These guidelines should specify when AI can be used, for what types of tasks, and how its involvement should be disclosed.
- Consistent Communication: Policies alone are not enough; they must be communicated consistently and clearly to all employees and, critically, to all levels of management. Managers need training on how to interpret and apply these policies, and how to evaluate AI-assisted work fairly.
- Invest in AI Literacy: Both employees and supervisors need to be educated on the capabilities, limitations, and ethical considerations of AI. This literacy can demystify AI, reduce apprehension, and foster a more informed approach to its integration.
- Emphasize Ethical AI Use: Policies should be anchored in ethical principles, stressing accountability, transparency, fairness, and the responsible handling of data.
These recommendations aim to bridge the gap between technological advancement and human perception, ensuring that the integration of AI enhances rather than erodes the foundational element of trust in the workplace.
The Path Forward: Building Trust in an AI-Driven Enterprise
The study does not posit that every employee using ChatGPT will invariably face lower trust or diminished promotion prospects. Instead, it serves as a critical warning, highlighting how managerial perceptions of AI use can significantly influence workplace evaluations, and, more optimistically, how explicit organizational support can fundamentally alter those perceptions.
For organizations striving to harness the full potential of AI, the path forward involves a multifaceted strategy:
- Strategic Policy Development: Moving beyond reactive bans to proactive, enabling policies that outline ethical, transparent, and productive AI use. These policies should be living documents, subject to regular review and adaptation as AI technology evolves.
- Fostering Open Dialogue: Creating channels for employees to openly discuss their AI use, share best practices, and raise concerns without fear of reprisal. This involves cultivating psychological safety within teams.
- Investing in AI Literacy and Training: Providing comprehensive training programs for both employees and managers. Employees need to learn how to use AI effectively and ethically, while managers need to learn how to supervise AI-augmented teams, evaluate AI-assisted work, and build trust in a hybrid human-AI environment.
- Reaffirming Human Value and Oversight: Continuously emphasizing that AI is a tool to augment human capabilities, not replace them. Stressing the importance of human critical thinking, creativity, and ethical judgment in overseeing and validating AI outputs.
In an era increasingly defined by artificial intelligence, the cultivation and maintenance of trust will be paramount for organizational success. This groundbreaking research from The Leadership Quarterly serves as a clarion call for leaders to proactively shape a future where AI and human ingenuity can coexist and thrive, built on a foundation of transparency, clarity, and unwavering trust.
