# Why So Many Workers Are Hiding How Much AI Is Actually Helping Them > The anxiety around AI at work may have less to do with learning the tools and more to do with what happens once a manager sees how much faster the job gets done. - Source: Morning Week - Canonical URL: https://morningweek.com/article/why-so-many-workers-are-hiding-how-much-ai-is-actually-helping-them - Author: Morning Week Newsroom - Section: Tech - Published: 2026-10-14T15:59:04.886Z - Updated: 2026-10-14T15:59:04.886Z - Tags: AI, future of work, job security, labor unions, workplace, automation, Norway --- The usual story about AI anxiety at work is that people are afraid of falling behind, worried they will not master the tools fast enough to keep up. A growing body of research points to a stranger, more uncomfortable version of the fear: workers who have already mastered the tools and are now quietly deciding how much of that mastery to let their bosses see. ## The incentive to stay quiet A survey of 6,000 computer-based workers by the workplace search company Glean found that 36 percent admitted to hiding just how much AI was helping them get their work done. A separate 2024 survey from Microsoft and LinkedIn found that more than half of workers using AI on important tasks worried that doing so openly would make them look replaceable. The pattern is consistent across both studies: the better the tool works, the more reason an employee has to downplay it. ## An old problem with a new name This behavior is not new, even if the technology is. In 1911, the engineer Frederick Winslow Taylor documented a practice he called soldiering, workers deliberately holding back their true productivity because they had learned that showing employers what they were capable of tended to be rewarded with lower pay rates or higher quotas, not recognition. More than a century later, the logic has resurfaced with AI playing the role that faster machinery once did. If a tool lets someone finish in three hours what used to take a full day, revealing that gap honestly can read less like good news and more like a job posting for a smaller headcount. ## Why the response differs by country Research into this dynamic has drawn a notable contrast between workplaces with strong collective bargaining and those without it. In Norway, where roughly 52.6 percent of employees belonged to a union as of 2025, workers have institutional channels to negotiate how the gains from new technology get shared. In the United States, where union membership sits at about 10 percent overall and just 5.9 percent in the private sector, most workers facing the same AI-driven efficiency gains have no equivalent seat at the table, and little choice but to manage the risk individually by staying quiet. The fear is not that AI will not work. The fear is that it will, and that whoever notices first will decide the job no longer needs to exist. ## A different model, built on disclosure instead of concealment Norway has been testing an alternative. A national agreement set to take effect in 2029 brings unions and employers together specifically to manage AI adoption, with the explicit goal of ensuring workers have a voice in how the technology is rolled out and a share in whatever benefits it produces, rather than simply absorbing the downside through job losses or unpaid speed-ups. Researchers studying the problem have floated a simpler version companies anywhere could try without waiting for national policy: written guarantees that a job will not be cut purely as a result of an AI adoption trial, paired with a joint decision between workers and management about what happens to the time that gets freed up, whether that means new projects, reduced hours, or something else negotiated rather than imposed. The idea is to remove the reason to hide the gains in the first place, so that honesty about what the tools can do stops being a professional risk. Until something like that becomes standard, the incentive structure many workers face stays the same as it was in Taylor's day, just with a chatbot instead of a stopwatch. The faster and more capable the tool, the more useful it can be to quietly keep that fact to oneself. --- Originally published by Morning Week. Free to cite with attribution and a link to https://morningweek.com/article/why-so-many-workers-are-hiding-how-much-ai-is-actually-helping-them.