Impostor Syndrome Myths Are Falling Apart, Just as AI Hands Us a New One
Research now shows impostor thoughts come and go and aren't always harmful. AI brings a different kind of self-doubt. Here is how to tell a real skill gap from noise.
If you write software for a living, you probably remember the first time a coding assistant produced, faster than you could finish reading the ticket, the function you had planned to spend the afternoon on. It compiled. The tests passed. And under the relief sat a smaller, colder feeling that was hard to name: if a tool can do that, what exactly have they been paying me for?
People are calling that feeling impostor syndrome. I think it is something slightly different, and the difference matters, because the research on classic impostor feelings has shifted a lot in the last few years. Some of what most of us believe about impostor syndrome turns out to be wrong. And the new, AI-shaped doubt needs its own handling.
What the Research Now Says About Impostor Feelings
The idea goes back to 1978, when psychologists Pauline Clance and Suzanne Imes described high-achieving women who were convinced their success was a fluke. They called it the impostor phenomenon. It was never a clinical diagnosis. The word "syndrome" came later, and it quietly made the experience sound like a condition you have rather than a thought you sometimes think.
Basima Tewfik, who studies work and organizations at MIT Sloan, prefers a narrower term: workplace impostor thoughts, meaning the belief that other people overestimate your competence. In a recent review in the Academy of Management Annals, written with Jeremy Yip of Georgetown and Sean Martin of the University of Virginia, she took apart several assumptions that had hardened into folk wisdom.
- It is not a permanent trait. Impostor thoughts come and go. You can have them in one meeting and not in the next.
- It is not mainly a women's experience. Across studies, men and women report it at similar rates.
- It is not uniformly harmful. In some settings, people who have these thoughts actually do better, at least on the interpersonal side of work.
- The damage story is less proven than it sounds. Many of the claimed ways impostor feelings hurt people have little solid data behind them.
That third point comes from Tewfik's own earlier work. In a 2022 paper in the Academy of Management Journal, built on four studies with more than 3,600 participants, employees who had impostor thoughts more often were rated by others as more interpersonally effective. The thoughts did not come at the cost of their performance, either. Her explanation is that feeling you might be found out turns your attention outward, toward other people. In practice, I imagine that looks like listening harder, asking more questions, checking in more often. You compensate.
None of this means impostor feelings are pleasant, or that they never tip into something heavier. It means the experience is more like weather than climate, and some of it is doing useful work.
Productive Self-Doubt Versus Paralyzing Impostor Feelings
So the practical question is not "how do I get rid of this?" It is "which kind of doubt am I having right now?"
| Productive self-doubt | Paralyzing impostor feelings |
|---|---|
| Specific: "I have never load-tested a service this size." | Global: "I don't belong on this team." |
| Points toward an action | Points toward hiding or avoiding |
| Changes when new evidence arrives | Survives every piece of evidence |
| Shrinks with practice | Moves on to the next achievement |
| Feels uncomfortable | Feels like shame |
The fourth row is the tell I trust most. Real skill gaps close. You practice, you get feedback, and the doubt about that particular thing gets quieter. Impostor noise just moves house. Get the promotion and it settles into the new role. Ship the project and it decides the project must have been easy.
Heartfulness meditation has a gentle instruction for thoughts that keep arriving during practice: don't fight them, and don't follow them. Let them be there, and bring your attention back to the heart. I have found impostor noise responds to that treatment far better than to argument, because it was never persuaded by evidence in the first place.
The New One: Competence Anxiety in the Age of AI
Classic impostor thoughts have a particular shape. People think I am better than I am. The standard stays still, and you doubt that you meet it.
AI-driven competence anxiety has a different shape, because the standard itself is moving. Work you were quietly proud of, such as clean boilerplate, a solid first draft, a crisp summary of a long report, can now be produced in seconds by a tool anyone can open. Your skill did not get worse. The value placed on it shifted, and it shifted at a speed most careers have never had to absorb.
Here is the mechanism as I understand it. Your professional self-concept is a summary built slowly out of years of evidence: I am the one who can untangle this codebase, who writes the clear memo, who knows the tax rules cold. It updates slowly on purpose, which is healthy. You would not want one bad meeting to rewrite who you are. But when the way work gets judged changes within a couple of years, that summary goes stale before it has time to refresh. For engineers, it is like reading from a cache after the database underneath has been migrated. Every read returns something that used to be true.
That gap produces a feeling that resembles impostor syndrome but isn't quite it. Classic impostor thoughts are usually miscalibrated: the evidence says you are doing fine, and the feeling refuses to believe it. AI anxiety is often partly calibrated. Some tasks you owned really have been absorbed. So reassurance alone ("you're great, don't worry") is as useless here as telling someone their job isn't changing when it plainly is.
What helps is separating the strands. Some of the anxiety points at a genuine gap, usually a newer skill: framing a problem precisely enough that a tool can help with it, judging output you did not write, noticing when a confident answer is wrong. Those gaps are learnable, and they close like any other. The rest is noise, particularly the sense that because a tool can now do part of your work, you were never really good at it. That conclusion does not follow. Spreadsheets did not retroactively turn accountants into frauds.
Should You Tell People You Feel Like an Impostor?
The usual advice is to talk about it. Name the feeling and you'll discover everyone has it. Among friends that is often true, and a genuine relief. At work, the evidence suggests more care.
In a study published in November 2025 in Scholarship of Teaching and Learning in Psychology, researchers at Colorado State University showed students two nearly identical descriptions of a tenured professor. In one version, the professor disclosed impostor feelings. Students rated that professor as less experienced and guessed they earned about $10,000 less. The same professor was rated just as likeable, and students were just as willing to enroll in the class.
It is one study, with students judging a hypothetical person, so I would not build a rule on it. But it matches something many of us have sensed: disclosure can buy warmth while costing perceived competence, and in settings where you are being evaluated, competence often counts for more. The decision is less "tell or don't tell" than who, when and how.
- Tell peers and mentors before evaluators. The person deciding your promotion needs to hear your plan more than your fear.
- Translate global doubt into a specific gap. "I feel like a fraud with these AI tools" lands very differently from "I'm still building a reliable way to review generated code, and here's what I'm trying."
- Consider timing. "I doubted myself going into that launch," said after it went well, reads as honesty. The same sentence on the morning of the launch can read as a warning.
A Two-Column Worksheet: Skill Gap or Impostor Noise?
This works best on one concrete situation, not your whole career. Pick something in the next few weeks that makes your stomach tighten. Write it at the top of a page. Then sort every worried thought about it into one of two columns.
Column one, evidence-based skill gaps. A thought belongs here if it is specific, if you can point to real evidence (feedback, a failed attempt, a requirement you have not yet met), and if you can name a step that would shrink it.
Column two, impostor noise. A thought belongs here if it is about who you are rather than what you can do, if no realistic evidence could settle it, or if it has survived earlier successes untouched.
A worked example
Situation: presenting a customer-churn analysis to leadership next week. About half the analysis code was drafted with an AI assistant.
| Evidence-based skill gap | Impostor noise |
|---|---|
| I have not checked the join logic the assistant wrote against a sample I verified by hand. | "Someone is going to realize I didn't really do this analysis." |
| Feedback on my last deck said the charts needed clearer takeaways. | "A real analyst wouldn't have needed a tool for this." |
| I can't yet explain why this clustering method was chosen over simpler ones. | "I've just been lucky so far." |
| Next steps: reconcile the output against 50 hand-checked rows; ask one colleague to read the deck only for clarity; read up on the method or swap to one I can defend. | What to do with these: notice them, label them as noise, and refuse to spend an hour arguing with them. |
Your turn
Situation: ____________________________
| Evidence-based skill gap | Impostor noise |
|---|---|
| 1. | 1. |
| 2. | 2. |
| 3. | 3. |
| One next step for each item: | Label it, and get back to column one. |
Two things usually happen when people do this honestly. The left column turns out shorter than the anxiety suggested, and every item in it has a next step. The right column is longer, and none of its items has one. That is the point. You were never going to argue your way out of those thoughts. You can only recognize them and return to the work on the left.
Frequently Asked Questions
Is impostor syndrome a mental health diagnosis? No. It does not appear in the DSM, and the researchers who first described it called it a phenomenon, not a disorder. If your self-doubt comes with persistent low mood, anxiety that disrupts your sleep, or a sense of dread that follows you home, it is worth talking to a mental health professional. That is not a failure of mindset.
Can impostor thoughts actually be useful? Sometimes. Tewfik's research found that people who have them more often tend to be rated as more interpersonally effective, likely because they pay closer attention to others. The trouble starts when doubt turns into avoidance: not applying, not speaking up, not shipping.
How is AI anxiety different from ordinary impostor syndrome? Classic impostor thoughts usually contradict the evidence; you are doing well and cannot believe it. AI anxiety often contains some accurate information, because the standards for certain tasks really are changing. That is why sorting real gaps from noise works better than blanket reassurance.
Should I hide that I use AI tools at work? Follow your workplace's policies first. Beyond that, my own view is that being plain about how you use a tool, and how you check its output, holds up better over time than either hiding it or apologizing for it. The checking is the part that is still unmistakably yours.
The tools will keep getting better at tasks. What stays with you is the judgment about which tasks matter, and the honesty to know, on any given Tuesday, which of your doubts deserve an answer and which only deserve to be noticed.