Remember when bosses made employees use AI? Now they regret it

Managers spent years urging employees to use AI, even as studies found no noticeable productivity gains. Now some companies are banning AI from work messages: managers are receiving a stream of meaningless, model-generated text. A study in The Leadership Quarterly found that using ChatGPT can reduce an employer’s trust and an employee’s chances of promotion. In experiments, the likelihood that a manager would trust an employee’s report fell from 68% to 24% when the manager knew the employee had used AI.

Workers were encouraged to use AI. Now managers trust them less

Work messages instead of benefits

Business Insider reports that an executive at a technology recruiting firm began receiving Slack messages that were clearly written with AI. In response, she imposed a strict ban on using such tools in internal communications.

Earlier, Inc reported that an anonymous executive threatened to fire the next employee who sent him unedited text from ChatGPT. These cases suggest that widespread AI use may not improve office work, but instead add meaningless messages to it.

AI use reduces trust

The study, published in The Leadership Quarterly, involved a team led by Samuel Meyers, a doctoral student at the Hebrew University of Jerusalem. In a series of experiments, managers were asked to decide whether they trusted an employee’s report. They were told whether the employee had used AI, but were not asked to assess the report’s quality.

Managers trusted employees who had not used AI in 68% of cases. When an employee had used AI, the figure fell to 24%.

Study co-author Yohanan Bigman, a senior lecturer at the Hebrew University of Jerusalem, said AI changes not only work processes but also how people judge their colleagues. An employee may use AI to do better work, faster, yet still lose a manager’s trust simply because they used the technology, he said. Bigman added that adopting AI is not just a technological challenge, but a psychological and managerial one as well.

Costs and uncertain returns

Managers have also found that AI agents and coding tools increase costs and hurt companies’ financial results. In some cases, businesses have had to revise plans and cut spending.

Researchers have yet to find significant benefits from using AI at work. Surveys show that the vast majority of managers have seen no increase in productivity. This is happening as labor-market conditions worsen, regardless of whether AI improves efficiency.

Employees face a choice: follow management’s instructions and use AI, or avoid it to preserve trust. Meyers and his colleagues worry that this tension will lead to “shadow AI”: the concealed use of tools at work, which could further harm employees’ career prospects.

In Dataist’s view, companies must consider not only AI’s potential effects but also how its use affects trust between managers and employees. The study highlights a risk: the mere fact that a worker turned to AI may affect how they are judged, even when the quality of their work has not been assessed.