Capital Wealth
TUE CLOSE · SEP 15   S&P 500 7,585.73 ▼0.45%  ·  DJIA 52,093.11 ▼0.63%  ·  NASDAQ 25,981.57 ▼0.78%  ·  10-YR 4.996%  ·  2-YR 4.673%  ·  WTI $106.05 ▲4.6%  ·  GOLD $4,337.00 ▼0.3%  ·  VIX 17.42 ▲1.9%
Journal Report · AI at Work

Think a Co-Worker Is Secretly Using AI? Your Hunch Is Basically a Coin Flip

A New York University psychologist says fuzzy AI rules are souring trust between colleagues. Most of us can’t actually tell who’s using it.

By Sean Anees Saifi · Capital Wealth · Published Tuesday, September 15, 2026 · Source: The Wall Street Journal, September 14, 2026 edition (Journal Report: Artificial Intelligence)
Key Points
40%
workers using AI where they weren’t supposed to
57%
admitted claiming AI-made content as their own work
50–54%
average accuracy at detecting AI content
Office workers sit at computer desks in a modern high-rise office with floor-to-ceiling windows overlooking a city at dusk.
Unclear AI rules leave co-workers guessing who leaned on a chatbot, and the guesses are barely better than chance. Tessa West says the company, not the cubicle detective, should settle it.
In one line: Workplace AI suspicion is mostly guesswork, so clear company rules beat private hunches, and your own reputation is worth guarding while you wait for them.

The email is long, monotonous and oddly impersonal. Your co-worker didn’t write that, you decide. A chatbot did. Tessa West, a psychology professor at New York University, has news for the detective in you: you’re probably guessing. In a Journal Report essay, she argues generative AI is eroding trust between co-workers. A big reason is that most organizations have no clear rules for how employees should use it.

A coin-flip hunch

The suspicion isn’t crazy. In one study of 30,000 employees, about 40% used AI at work even when they weren’t supposed to. KPMG found that about 57% of 48,000 employees admitted trying to present AI content as their own.

Spotting it is another matter. On average, people identify AI content correctly just 50% to 54% of the time, which is chance, more or less. They lean on misleading cues, like how long or impersonal an email is. When a hypothetical worker named Sara sizes up her colleague Scott, she is relying on “a coin flip, and one she doesn’t know she’s made.”

The penalty for being suspected is steep. We see AI users as lazier and less competent and their emails as less genuine, all while hiding our own use. Since nobody enjoys a confrontation, the hunch often just simmers into resentment.

Your name is an asset

The fixes she proposes are aimed at employers. Form groups of people from different backgrounds to discuss how AI is really used. Survey workers anonymously. Publish a guidebook that marks the black, white and gray areas, including what people are expected to disclose. Then report on AI use in regular company updates, so nobody has to guess.

Our take, for anyone in the late innings of a career: don’t play detective. Your reputation is an asset built over decades, and a wrong accusation can dent it faster than any shortcut a colleague takes. Learn the rules where you work, follow the disclosure norms yourself, and let the policy settle the rest.

Nobody shops for an umbrella once it’s already pouring. If your last working years are part of the retirement math, bring a recent statement and give us fifteen minutes to look it over.

What It Means For Your Portfolio

Hold — know the rules, skip the accusation

Don’t act on a hunch about a colleague’s AI use; learn your employer’s rules, follow its disclosure norms, and protect a reputation that took decades to build.

A professional reputation behaves like any long-held asset: slow to build, quick to damage. In the final working years it can shape raises, references and any part-time work that bridges you into retirement. Know the rules before acting, disclose your own AI use where disclosure is expected, and raise concerns through proper channels. These are general planning principles, not advice for anyone in particular.

Nothing in the Capital Wealth portfolios changes because of this story; we’re still neutral on stocks and buying nothing new ahead of the Fed decision on Wednesday. The household move is a habit: this week, read the AI policy at work, and if there isn’t one, ask your manager what’s expected.

Book a 15-Minute Review → Back to Edition No. 170 →