Skip to main content
EDIT

Why Are Employees Reluctant To Disclose AI Use to Their Bosses?

Why Are Employees Reluctant To Disclose AI Use to Their Bosses?

A new study from Associate Professor Eric Anicich and PhD student Jeslyn Brouwers explores how organizational mistrust discourages workers from being transparent with their AI use.

by: Alexander Bernard
08.28.26
A human and robot working on different computers

[iStock Photo]

Featured

Stay Informed + Stay Connected

MARSHALL MONTHLY BRINGS YOU ESSENTIAL NEWS AND EVENTS FROM FACULTY, STUDENTS, AND ALUMNI.
SIGN UP

Countless employees across the country are experimenting with artificial intelligence (AI) to streamline workflows and increase productivity. Yet, many are keeping these solutions to themselves.

According to a study conducted by two researchers at the USC Marshall School of Business, many workers are reluctant to disclose how they’ve incorporated AI into their professional routines. Eric Anicich, associate professor of management and organization, and Jeslyn Brouwers, a Marshall PhD student, surveyed 604 U.S.-based employees who reported using AI at work daily or multiple times per day.

While four out of five respondents believe sharing their techniques would benefit their team, nearly one out of three withheld “AI-related knowledge, workflows, or techniques from coworkers or employers.”

Whereas some employees’ concern centered around job loss and internal competition, much of their hesitation derived from mistrust of their organization.

“What separated the sharers from the hoarders was trust in the organization,” Anicich explained. “People in the bottom quarter on trust were about four times as likely to have withheld something as people in the top quarter. It boils down to psychological safety, which is about whether you feel comfortable sharing what you know without worrying you'll be ridiculed or punished for it.”

Although some companies have taken steps to encourage knowledge sharing, employees remain skeptical of how leadership would react to full disclosure. In fact, the study points to three factors that contributed to this lack of trust.

First, employees were concerned their reputation would be tarnished if they admitted to using the technology. Anicich says other research has validated their anxiety.

“Absent a clear signal that their organization values AI experimentation, withholding is the rational move — and a growing body of research suggests employees are right to worry,” Anicich explained. “Indeed, research shows that learning that a person used AI can lead observers to attribute less competence, motivation, effort, creativity, authenticity, or trustworthiness to that person and, in some contexts, to discount the value of the resulting work.”

Second, respondents worried their supervisors would view their improved efficiency as an excuse to give them more work or even to make them a “spokesperson” for AI within the company. In interviews, Brouwers and Anicich heard from multiple CEOs who sought to assuage these concerns while still rewarding transparency.

“In some of the CEO interviews, they highlighted that something really important for them was to validate the hesitation of [being the] spokesperson for AI from now on,” Brouwers explained. “[They try to communicate:] We’re not changing your whole job title.”

Finally, by demonstrating AI’s proficiency in their job responsibilities, employees were anxious they were proving that the technology could eventually replace them.

It’s a very different story to send a company memo out saying that we are now a culture where we share knowledge and we share innovations and share all of the secret things that you’re doing versus showing that [philosophy].

— Jeslyn Brouwers

Marshall PhD student

“Right now, knowing how to actually get results out of these tools is a differentiator. It’s what makes someone hard to swap out,” Anicich said. “But the moment that knowledge comes out of your head and becomes something the company has, you’re not as distinctive anymore. A colleague can do what you do. And a few years out, the tools themselves might be able to.”

Organizations might not get a second chance at building trust. Across their surveys, the co-authors observed that employees who divulged their AI use could tell within the first thirty seconds whether their supervisors were supportive.

“That’s when the employee is [answering the questions]: Am I punished for this? Will I be given more work?,” Brouwers explained. “If [these hesitations] are realized right after they share the knowledge, then that knowledge will not be shared in the future.”

Anicich and Brouwers explain that the withholding of useful AI techniques can negatively affect a company’s immediate output and long-term potential. Not only are efficiency gains limited to a single employee, but Anicich points out that business leaders may also be making decisions based on an inaccurate perception of their organization’s realities, including relative workload and staff performance.

“If your best analyst is fast because of a method nobody has seen, you can’t train anyone else on it, and you don’t really know what you're paying for,” Anicich said.

Rebuilding organizational trust is far from simple. Anicich and Brouwers stress that organizations must make cultural changes as well as structural ones. It’s not enough to release a policy that claims to support AI use.

“It’s a very different story to send a company memo out saying that we are now a culture where we share knowledge and we share innovations and share all of the secret things that you’re doing versus showing that [philosophy],” Brouwers said.

Workers must see tangible benefits for sharing their innovations. According to the researchers, leaders must model this support in day-to-day interactions by recognizing and rewarding innovation — without making the employee a “spokesperson” for the technology.

Possible rewards might include credit in performance reviews for methods that others adopt, protected time to keep experimenting, and a share of the gains once a workflow is in wider use,” Anicich and Brouwers wrote in Harvard Business Review.

By incentivizing a culture of experimentation and risk-taking, the co-authors believe companies can rebuild trust with their employees, encourage a psychologically safe and collaborative workforce, and more easily realize their organization’s potential.

“Culturally, managers need to remain curious and willing to reward productive experimentation with AI rather than treating it by default as corner-cutting or deviance,” Anicich said. “Over time, employees come to learn that this kind of experimentation, and the productivity gains that come from it, are valued and rewarded by the people in a position to evaluate their performance. And then they become more willing to both experiment and share the results with others in the organization.”

RELATED