AI: Your Dirty Little Secret?

Last week, a colleague pulled me aside after a meeting. Lowered her voice. Looked both ways like she was about to confess to a felony.

“Mollie,” she whispered. “I used ChatGPT to do all that research.”

She said it the way someone might admit to putting ketchup on a steak. Shameful. Illicit. Deserving of judgment.

I wanted to shake her. Instead, I laughed and said, “Sweetie, I used AI every day, multiple times a day. Welcome to the club.”

The relief on her face? Heartbreaking. And infuriating. Because here we are, in 2026, and smart, accomplished professionals are still treating AI use like a dirty little secret. Like something you do in private and never, ever talk about at work.

We need to talk about why. And then we need to stop.

Here’s the thing: I get it. I really do.

When I first started using AI tools, I felt it too. That nagging voice whispering, Are you cheating? Is this real work? Will people take me less seriously if they knew?

(For the record, that voice sounded suspiciously like every imposter syndrome gremlin I’ve battled for 25 years in public health. Same gremlin, new outfit.)

The shame comes from everywhere. Academic culture tells us that every word must be painstakingly original. Professional culture rewards the appearance of effortless competence. And our own egos whisper that needing help means we’re not good enough.

So we hide it. We close the ChatGPT tab when someone walks by. We say “I drafted” when we mean “I prompted and then heavily edited.” We pretend the beautifully formatted report materialized from sheer willpower and three cups of coffee.

Sound familiar?

A recent survey from Microsoft found that 75% of knowledge workers are already using AI at work. But here’s the kicker: 52% of those users said they were reluctant to admit it for fear of being seen as replaceable or lazy.

Seventy-five percent are using it. More than half won’t admit it.

That’s not a productivity tool. That’s a speakeasy.

Now, if this were just about feelings, I’d say work through it on your own time. But the AI shame spiral has real consequences, especially in public health.

It slows adoption where we need it most. When nobody talks about using AI, nobody learns from each other. Your colleague who figured out how to use Claude to synthesize 200 community health assessments in an afternoon? She’s keeping that magic to herself because she’s afraid someone will think she cut corners. Meanwhile, the rest of the team is drowning in documents, doing it “the real way.”

It prevents honest conversations about limitations. If you can’t admit you’re using AI, you definitely can’t say, “Hey, the AI gave me a biased output and here’s how I caught it.” We need those conversations. The only way we learn to use these tools responsibly is by talking openly about what goes right AND what goes spectacularly wrong.

It creates a two-tier system. The people who can afford to quietly use AI (folks with access, tech literacy, and the right devices) get ahead. Everyone else falls behind. And nobody names what’s happening because the whole thing is shrouded in secrecy.

(Sounds a lot like every other equity issue in public health, doesn’t it? The tools that could level the playing field instead widen the gap, because access and knowledge aren’t distributed equally. Do better, y’all.)

It makes us worse at our jobs. I’m serious. When you’re sneaking AI use instead of integrating it thoughtfully into your workflow, you’re not building good practices. You’re not documenting your process. You’re not teaching your team. You’re just... hiding. And hiding is exhausting.

Let’s address the elephant in the room. The biggest source of AI shame is the belief that using AI means you didn’t do “real work.”

I want to push back on this. Hard.

Did you define the problem? That’s real work. Did you know what questions to ask? Real work. Did you evaluate the output critically? Real work. Did you edit, refine, and add your expertise? Very real work. Did you make judgment calls about what to include and what to cut? The realest work of all.

When I use AI to help draft a document, I’m not outsourcing my thinking. I’m outsourcing the part where I stare at a blank page for 45 minutes, write a mediocre first paragraph, delete it, write another one, delete that, and then go get more coffee.

The thinking? The strategy? The “should we frame this around health equity or economic impact” decision? The knowledge that this particular funder cares about community engagement metrics? That’s all me. AI doesn’t know any of that. AI doesn’t know my community. AI doesn’t know my field the way I know my field.

Using a calculator doesn’t mean you can’t do math. Using spell check doesn’t mean you can’t write. Using AI doesn’t mean you can’t think.

It means you’re being smart about where you spend your cognitive energy. And in public health, where we are chronically understaffed, underfunded, and overwhelmed? That’s not laziness. That’s survival.

So how do we break the shame cycle? We start by being honest. Loudly.

Here’s what I’ve been doing, and I’d love for you to try it too:

Name it in your work. I’ve started adding a line at the bottom of my documents: “AI tools were used in the drafting process. All content was reviewed, edited, and verified by the author.” It’s simple. It’s honest. And every time I include it, it normalizes the practice for whoever reads it next.

Talk about it in meetings. “I used Claude to pull together the first draft of this framework, and then I spent two hours reshaping it.” That’s not a confession. That’s a workflow update. Say it like you’d say “I used Excel for the data analysis.” Because it’s the same thing.

Share your prompts. This is my favorite one. When someone compliments a piece of work, tell them how you made it. “Thanks! Here’s the prompt I used to get started. Want to try it for your next report?” You become the generous colleague, not the AI cheater.

Be honest about failures too. “I tried using AI for the community needs assessment summary and it completely missed the cultural context. Had to rewrite the whole thing.” These stories are gold. They teach people what AI can’t do, which is just as important as what it can.

Create safe spaces for your team. If you’re a supervisor, say it first. “I use AI for X, Y, and Z. I’d love to hear how you all are using it, or if you want to learn.” Watch the floodgates open.

I can’t write about AI shame without talking about who gets to feel ashamed and who doesn’t.

Let’s be honest: the tech industry is not hiding its AI use. Silicon Valley is not whispering about it in hallways. They’re building entire companies around it, raising billions, and putting “AI-powered” on everything from toasters to toilet seats.

The people hiding? They’re in public health. In education. In nonprofits. In government. In exactly the sectors that could benefit most from these tools but where the culture of “doing it the hard way” is most deeply entrenched.

And within those sectors, it’s often the people with the least institutional power who feel the most pressure to hide. The early-career professional afraid her boss will think she’s not pulling her weight. The community health worker who doesn’t want to seem like she doesn’t know her own community. The grant writer worried the funder will reject the proposal if they knew AI touched it.

Meanwhile, the consulting firm your organization hired at $300 an hour? They’re absolutely using AI. They’re just not telling you either.

The shame isn’t equally distributed. It never is.

If we want AI to be a tool for equity rather than another mechanism of inequity, we have to democratize not just access but permission. Permission to use it. Permission to talk about it. Permission to be imperfect at it while you learn.

Here’s what I want you to do. (Consider this your formal permission slip, signed by a DrPH who curses at her computer regularly and once fell down the stairs “gracefully.”)

This week, tell one person you use AI. Just one. A colleague you trust. Say it casually. See what happens. I bet they’ll say, “Oh thank God, me too.”

Next week, share a prompt. Send a colleague a prompt that helped you with something. “Hey, I used this to draft talking points for the board meeting, feel free to adapt it.” Watch their face light up.

This month, add a transparency line to one document. Just one. Put it at the bottom. See if the world ends. (Spoiler: it won’t.)

And if anyone gives you grief? Tell them that hiding AI use doesn’t make anyone smarter. It just makes everyone slower, more stressed, and less honest.

Public health has never had the luxury of doing things the hard way just for the sake of appearances. We’re too busy. We’re too underfunded. We have too many communities counting on us.

If AI can help us do better work, faster, with more time for the human stuff that actually matters (listening to communities, building trust, showing up when it counts), then using it isn’t something to be ashamed of.

It’s something to be proud of.

Now stop whispering about it. Start talking.

And do me a favor? Share this with that colleague who’s been hiding their ChatGPT tab. They need to hear it.

Leave a comment

PS: I used Claude to help me draft and structure this post. The rant? The frustration? The conviction that we need to do better? All mine. Claude just helped me stop staring at a blank page. See how easy that was?