When people use AI quietly at work, what everyday tips and tricks stop spreading between coworkers?

This explores what a workplace loses in everyday, over-the-shoulder learning when people use GenAI but don't tell their colleagues. The corpus says the hiding is common and deliberate. One reason surprises people: hiding AI use isn't only about avoiding stigma, it's also a way to look like an expert. In interviews with knowledge workers, people removed signs of GenAI from their work because polished output that seems to come from their own skill earns credit Why do knowledge workers hide signs of using GenAI?. At scale, a 1,250-person interview study found that about 70% of workers hid or played down their AI use even though most said it saved them real time Why do workers hide productivity gains from AI use?. People have reason to expect a penalty. Across four experiments, AI users predicted colleagues would rate them as less competent and less diligent, and they said they were less willing to tell managers and coworkers Do people fear judgment when they use AI at work?. Human raters did in fact mark down work when AI use was disclosed Do LLM raters show hidden demographic preferences that disclosure erases?.

The clearest loss is know-how about when and how to use these tools at all. Nielsen Norman Group found that the main barrier to using AI chat for finding information isn't resistance. It's that people don't know when the tool helps or how to use it well Does generative AI chat actually replace traditional search?. That kind of knowledge usually spreads by watching a coworker do it. When use goes underground, each person has to work it out alone, and the interview study links this directly to less informal knowledge-sharing between peers and to workplaces where secrecy about AI becomes normal Why do knowledge workers hide signs of using GenAI?.

A less obvious loss: certain information disappears from the finished work itself. In the 1,250 interviews, workers carefully kept cues tied to their identity, such as their voice and their name on the work. But they let cues about effort, attention, and uncertainty disappear into the deliverable Which workplace cues survive AI mediation and which disappear?. Those are exactly the signals a junior colleague learns from: where the hard parts were, what the author wasn't sure about, and which parts deserve a second look. A seamless document teaches nothing about how it was made. When writers who shared an editor could see each other's AI prompts, most wanted that visibility. It helped them understand a collaborator's thinking and check AI-written passages, though some found full sharing intrusive Do writers want to see each other's AI prompts in shared editors?.

This matters because the same tool can build skills or wear them down depending on how it's used. A study of Dutch knowledge workers found four outcomes: skills developed, maintained, eroded, or valued differently, depending on use patterns and on which tasks changed How does generative AI actually change worker skills?. If use is hidden, nobody can see which colleagues are learning with the tool and which are letting skills slip, so good patterns don't spread and bad ones aren't caught. Solo AI use is also a weaker way to learn than you might expect. In an 8-day field experiment, ChatGPT users learned less than Google searchers, especially on critical thinking, because the chat format steered them toward ready answers rather than exploring the wider topic Does ChatGPT harm informal learning compared to Google Search?.

A caveat on the evidence: the corpus shows clearly that people hide AI use, why they do it, and which cues vanish. None of these studies directly measures the learning that is lost. The connection is drawn from interviews and preference studies, not from tracking what teams fail to learn over time.