You know the feeling. You are mid-conversation with Claude, working through an email, a deck, a report, a tricky spreadsheet formula (whatever your version of “the work” is) and you hear footsteps. Your manager rounds the corner. Without thinking, your hand jumps to the keyboard and the window minimizes. By the time they are behind you, your screen is Outlook, a Word doc, and a polite-looking Teams tab. Nothing else.
Nobody told you to do that. You just did it.
That reflex is the thing worth talking about.
The quiet shame around using AI at work
A surprising number of professionals are doing this right now. Analysts cleaning up a quarterly summary, marketers drafting campaign copy, project managers turning a forty-minute meeting into action items, finance folks asking it to explain an Excel formula they would rather not admit they forgot. All of them are getting value out of it. A lot of them are hiding it.
Not because they are doing anything wrong. Somewhere along the way, “I used AI to draft this” started to feel like an admission rather than a method. People worry it will look like they cheated, that they are not “really” doing the work, that their value as an employee is somehow lower because part of the process was assisted. So they hide it, send the email, and let everyone assume it was done the old way.
That is a bad equilibrium. The output is better, the turnaround is faster, and nobody can talk about it openly. Which means nobody on the team can learn from each other, swap what is working, or get any better at it together.
Forward-thinking companies are doing the opposite
The companies that are getting this right are not subtle about it. They are openly telling their teams to use AI, paying for the licenses, sharing prompts internally, and treating fluency with these tools as a real skill rather than a guilty shortcut.
I happen to work at one of those companies, which is part of why this topic is on my mind. At Contemporary Managed Solutions, my boss Phil Hudson made the call early on to pay for our AI subscriptions out of the company’s pocket and treat staying on the cutting edge of these tools as part of the job, not a side hobby. The seats sit on our accounts already, waiting to be used. Nobody has to expense it or quietly hope it is allowed. That signal matters most when it comes from a leader who actually uses the tools, rather than one recycling a take from a podcast.
That changes the entire framing. The expectation is not that you justify using AI. It is that you use it well. You know when to lean on it, when to push back on its output, and when a draft needs to be rewritten end-to-end before it goes anywhere. Suddenly the conversation at work is not “did you use AI for this” but “what is your workflow, and is there a better one I should steal.”
That is the conversation you want to be having. And it only happens when leadership has signaled, clearly, that this is encouraged.
A tale of two CEOs
Picture two CEOs in the same industry. Same size company. Hiring from the same talent pool.
CEO A spots an employee using ChatGPT to draft a client email and reacts with quiet disappointment. Maybe they say it out loud, something like “you should be using your own critical thinking.” Or maybe they just let the look on their face deliver the message. Either way, the team learns the rule. Use the tool, but feel bad about it. Get the work done, but never admit how. Show me you can still do it the hard way, just in case.
CEO B sees the exact same thing and asks a completely different question: “what was the prompt? Did it get you all the way there, or did you have to rework it?” They are not policing the tool. They are paying attention to the workflow because they want the best version of it spreading across the team. They already accepted that AI is part of how the business runs. The only thing left to figure out is how well the team is using it.
Now run those two companies forward two years.
CEO A’s people are still using AI (they were never going to stop) but they are using it in private, on personal accounts, with no shared knowledge of what works. The best of them notice that nobody is openly building skills here, and they leave for somewhere that values it. CEO A is left wondering why turnover is up and why the proposals, reports, and decks coming out of the team feel slower and weaker than they used to.
CEO B’s people are openly comparing notes every week. The good prompts get shared. The bad outputs get torn apart in public. The team is visibly getting better at AI-assisted work, and the output (proposals, reports, customer responses, financial models, you name it) keeps getting sharper.
Same industry. Same size. Same starting talent. Wildly different trajectories. The only variable was how leadership treated the tool.
The “critical thinking” line, examined
The most common form of AI shaming arrives dressed up as a principled concern: “you should be using your own critical thinking, not letting AI do the thinking for you.” It sounds reasonable. Usually it is not.
Using AI well requires more critical thinking, not less. You read what it gave you, catch where it is wrong, push back, edit, rewrite, and own the final output that goes to a client or a board. The employee who blindly copy-pastes AI output is not thinking critically, and that produces bad work the same way any other shortcut does. The employee who uses AI as a sparring partner, pulls the best ideas from three drafts, and rewrites the rest is doing more thinking, not less. Conflating the two is not a defense of rigor. It is a way to sound principled while avoiding a tool the leader has not bothered to learn.
What “encouraging AI” actually looks like
It is not a slack message that says “feel free to use ChatGPT.” That is necessary but not sufficient. The behaviors that actually matter:
- Managers using it themselves, visibly. When the people running the team are open about how AI fits into their own emails, decks, and reports, the rest of the team stops worrying about being judged for it.
- No shaming in reviews. “Did you AI-generate this?” is the wrong question to ask about a memo, a proposal, or a presentation. “Is this clear, accurate, and useful?” is the right one. Work gets judged on the work.
- Treating AI fluency as a real skill. Hiring for it, training for it, talking about it in performance conversations. Pretending it does not belong on a resume in 2026 is its own kind of denial.
- Paying for the good tools. If the company expects AI-assisted output but will not buy the seats, employees end up pasting company information into personal accounts on personal devices, which is worse for everyone, including IT and legal.
None of this is radical. It is the same approach companies eventually took with email, the internet, smartphones, and video calls once those stopped being optional.
The controversial prediction
Here is the part I am willing to put my name on: the companies that embrace AI in their workflows (that encourage it, fund it, and refuse to make it taboo) are going to materially outperform the companies that treat it as something to be policed, hidden, or quietly tolerated.
Not by a small margin. By a meaningful one.
The reason is compounding. A team that is openly sharing prompts, comparing tools, and refining workflows together gets better at AI-assisted work every week. A team where everyone is using AI but pretending they are not gets none of that compounding. They are doing the same work twice (once with the tool, once with the cover story) and learning nothing collectively from either pass.
Multiply that across a few hundred employees and a couple of years, and the gap is not a productivity boost. It is a structural advantage. The “embrace it” companies will turn work around faster, write better proposals, respond to customers quicker, and hold onto the people who actually want to work this way. The “hide it” companies will lose those people first, then wonder where their pace went.
I do not think this is a prediction that ages badly. I think the only question is how long it takes for the laggards to notice.
What to do on Monday
If you are an employee hiding it: stop hiding it. Bring it up in your next one-on-one. Ask your manager directly how they want you to be using AI. The answer will tell you a lot about the company you work for, and either way you will be operating from a clearer position than you are now.
If you are a manager: be the one who says it out loud first. Tell your team you expect them to use AI, you want to hear about their workflows, and nobody is going to be penalized for being honest about how the work got done. Then mean it.
The minimize-the-window reflex is not a personal failing. It is a culture signal. And cultures change when someone decides they are done pretending.