TBSP (Too Busy, Summary Please): When AI makes someone faster at their job, telling anyone can feel risky. In one survey of about 6,000 workers, 36% of the people AI had helped said they hid how much it helped. They fear that a faster way to work can turn into a higher target or a smaller team. This week: why AI adoption at work stalls on trust more than skill, and what managers can do so people share what they learn.
If AI let you do your job in half the time, would you tell your manager? AI adoption at work has a problem that’s hard to measure, because the people involved keep it out of sight. Of the workers who say AI helps them do more and better work, 36% say they’ve hidden how much. That figure comes from a Glean survey of about 6,000 people in the U.S., U.K., and Australia.
The people hiding their gains have figured out the technology. What they can’t predict is how their employer will respond once it’s clear how much faster they’ve become.
The Fear Is Older Than AI
Information science professor Christoph Trattner traces the hiding back more than a century. In 1911, Frederick Winslow Taylor wrote about workers deliberately slowing down, a practice known as “soldiering.” Their reasoning was simple. If they produced more, employers might cut what they paid per piece, so working faster could mean more work for the same money. Taylor wanted to stamp out soldiering. Trattner’s point is that the fear behind it never went away.
The stakes look different today, but they work the same way. Trattner also points to a 2024 Microsoft and LinkedIn survey. More than half of people using AI at work worried that using it on important tasks would make them look replaceable. Someone who shows they can do their work much faster might get a higher target, more assignments, or a hard question about headcount. As Trattner puts it, “showing how much faster they are may be how the job gets smaller.”
The cost lands on the whole team. When one person keeps a faster process to themselves, nobody else learns it, and the manager never finds out what the tool can really do. Trattner’s conclusion is direct. “Today’s employers also need an answer for the person who finds a faster way to work.”
The Pitch Shapes the Use
Stanford researcher Arvind Karunakaran and his colleagues Roshni Raveendhran and Tami Kim found that the way a manager introduces AI shapes how much people use it. They interviewed 184 employees at a law firm, an ad agency, and an IT services firm. Then they spent nearly two years observing two matched divisions of a law firm that received the same AI drafting tool.
The only real difference was the pitch. One manager framed the tool as a productivity boost. The other framed it as a way to enrich the job. Paralegals in the second division used the tool 58% more and experimented with it 70% more. They also moved into legal research and case analysis. In the productivity-framed division, people used it only for the tasks they already had.
Making AI Adoption at Work Safe to Talk About
In Crucial Conversations, safety rests on two conditions. Mutual Purpose means people believe you care about their goals. Mutual Respect means they believe you care about them. Both studies point at the first condition. When the only goal people can see is more output for the company, hiding a shortcut or sticking to old tasks is a reasonable way to protect themselves.
People share what they learn when they trust that sharing won’t cost them. A few practices from the research help make the goal a shared one.
- Decide together where saved time goes. Trattner suggests letting employees help decide how saved time gets used. Managers can also talk through results with people before raising targets. Otherwise the saved hours fill up with more of the same, since work expands to fill the time it’s given.
- Put the promise in writing. Trattner’s proposal includes a written guarantee that nobody loses pay or a job during a trial. It also gives people a way to challenge a broken commitment.
- Lead with what the work could become. The paralegals who heard AI pitched as a way to enrich their jobs are the ones who stretched into new work.
- Protect time to explore. The successful division had formal training, weekly knowledge-sharing lunches, and weekly time set aside to experiment.
- Check what you measure. When paralegals used AI for deeper research, they handled fewer cases. Judging them by case volume alone would penalize the behavior the firm wanted.
If you’re the one who found the faster way, ask your manager directly what happens to the time you save. That conversation is easier before the shortcut becomes a secret.
Most AI adoption at work stalls on trust long before it stalls on skill. If you found a way to do your job in half the time, what would it take for you to tell your manager?
Leave a Reply