Most people who save time with AI at work keep quiet about it — what would make them comfortable admitting it?

This explores what conditions lead people to be open with coworkers about how they really use generative AI. The collection has much more on why people stay quiet than on what gets them talking. Taken together, though, those reasons suggest what openness would require. The basic pattern is striking. In a 1,250-person interview study, most workers and nearly all creatives said AI saved them time, yet about 70% hid or downplayed using it Why do workers hide productivity gains from AI use?. They cited workplace stigma, worries about professional identity and fear of being replaced. Experiments with more than 4,000 people found the same thing: AI users expect to be seen as less competent and less diligent, and they choose to tell managers and colleagues less Do people fear judgment when they use AI at work?.

The less obvious finding is that hiding isn't only defensive. Interviews with knowledge workers show that removing the traces of AI from your work can be a way of showing expertise Why do knowledge workers hide signs of using GenAI?. A polished document with no visible AI help says "I know this field." That matters for your question. If concealment pays off, simply reducing stigma may not be enough to get people sharing, because staying quiet still earns credit. The same study points to the cost: colleagues lose the informal, over-the-shoulder learning that usually spreads new working habits through a team.

The clearest evidence of willingness comes from tool design rather than workplace culture. When pairs of writers tried shared editors with different levels of AI visibility, most preferred seeing more of each other's prompts: when AI was used, how and where Do writers want to see each other's AI prompts in shared editors?. It helped them follow a collaborator's thinking and check AI-written text. Some found full visibility intrusive and made them self-conscious. The takeaway is that sharing comes more easily when the tool makes it the default, so nobody has to decide to confess. It also helps when the purpose is shared work rather than judging individuals.

Two other findings explain why this is hard to fix. Heavy AI users shift toward solo documentation work and add relatively little communication Does generative AI shift knowledge workers away from communication?. That leaves fewer natural moments to compare notes. And where use is still small and experimental, as among staff at Argonne National Laboratory, few people have settled practices worth describing How are national lab staff actually using generative AI?. Related research on sharing with AI itself offers a mirror image. People tell machines more than they tell other people because a machine doesn't judge them How do people decide what to share with AI systems?. Judgment is the barrier in both cases.

The gap is worth saying plainly. The collection doesn't yet test anything that makes colleagues more open, such as manager role-modeling, team norms or rewards for sharing. What it suggests is that openness depends on whether AI use reads as skill or as a shortcut. Research showing AI can rebalance which skills count, not just erode them How does generative AI actually change worker skills?, hints at the framing that could make sharing feel like showing expertise rather than admitting to a shortcut.