Most people who use AI at work hide it, and the research shows why: disclosure is penalised. In a KPMG and University of Melbourne study of about 48,000 people in forty-seven countries, 57% said they hide their AI use and present AI work as their own. Competitors who disclose and charge less will force the change.
Three reasons to keep quiet
In August we wrote that we built the perfect conditions for hiding AI use. The research since then shows three reasons people keep quiet.
There is no rule to follow. In the KPMG and Melbourne study, only 47% of employees had any AI training. The same study found 48% admitted using AI in ways that broke company policy. With no clear rule, the safest move is to use the tool and say nothing.
Your peers will think less of you. Jessica Reif, Richard Larrick and Jack Soll at Duke's Fuqua School ran four preregistered experiments. People who used AI were rated lazier, less competent and less diligent than people who got the same help from a colleague or worked unaided. Participants expected the penalty, and it applied anyway.
It makes your work look cheaper. In Microsoft's survey, 53% of AI users worried that using it on important work makes them look replaceable. If a draft took ten minutes, why does the invoice say four hours? Much professional pricing is a claim about effort, and AI breaks the link between effort and output.
An old accusation
Around 370 BC, Plato has Socrates tell of an Egyptian god who offers writing to a king as a gift for memory. The king refuses. People who rely on letters, he says, "will appear to be omniscient and will generally know nothing."
The usual story is generational: older professionals call AI a crutch for younger ones, as they did with Google and Wikipedia. Some people do use AI to fake competence. But the same charge met calculators and search engines, and each time the prestige of the old way was the first thing to go.
The Duke data does not fit the generational story. The penalty was the same at every age, and young colleagues mark each other down too. So it will not fade as older professionals retire.
The market will settle it
Professions could slowly update their norms, but a century of billable hours will not be undone by persuasion. Competition will undo it.
Clio's 2026 report on North American solo and small firms found that three quarters use AI. Fewer than a third report more revenue from it, compared with nearly 60% of large firms. The work got faster and the billing did not follow, so the client keeps the saved hours.
The debate about real expertise stops mattering once a competitor offers the same result for less. The bar association decides what counts as craft, and your client decides what to pay for. If you do not pick up the tool, a competitor will.
Why quiet use loses
First, you cannot price what you cannot admit. A firm that says "AI drafts, a partner checks every clause, and our fee reflects both" can compete on price and trust. A firm that hides its AI use gives itself away when it cuts its price.
Second, hidden use is unreviewed use. The same KPMG study found 57% of employees rely on AI output without checking its accuracy. Client contracts, patient notes and HR files go into those tools with no control. At a ten-person firm, nobody finds out until the client does.
What disclosure looks like
Disclosure has three parts, stated next to the work.
Say where you used AI. For example: "AI drafted the summary, and I rewrote the recommendation." One line on the work means nobody has to guess.
Say who checked it. Accountability is what the profession was defending all along. The publicised AI failures in law, such as invented citations, were punished because nobody reviewed the work. A named reviewer answers that objection.
Say what data the tool saw. The practical answer is to remove names and identifiers before the document goes to the model, and restore them locally afterwards. That is why we built promptShield, and it only helps people who admit they use AI.
Every post here carries the byline "drafted with Claude, reviewed and edited by hand." We rejected a version that put the model first, because it misstated who is accountable. Putting the model first also made us look worse, which is the penalty this article describes.
The age of disclosure
Nobody will announce the change. It will show up in a lower quote or a proposal that lists its method. Job ads will ask which tools you use and how you check them.
Firms that hide their AI use will be the slowest to build a new path for junior staff, who used to learn the craft on first drafts.
You do not have to wait for your profession. Tell one colleague how you actually work this week, and put the method next to the output. Competition will make disclosure normal either way, so you can get ahead of it or get caught out by it.
Common questions
How many employees hide their AI use?
In the KPMG and University of Melbourne study of about 48,000 people, 57% said they hide it. Microsoft and LinkedIn's 2024 Work Trend Index found 52% of AI users reluctant to admit it for important tasks. Two independent publishers reached the same result.
Why is disclosing AI use penalised?
Duke researchers found that colleagues judge AI users as lazier and less competent, at every age. People also fear it makes their work look cheaper.
What would make disclosure safe?
Disclosure is safe when clients pay for accountability rather than hours. Until then, attach the method: where the tool was used, who reviewed it and what data it saw. Pasting documents into ChatGPT covers the data half.