Client Says My Writing Is AI? What Freelancers Can Prove
Sally McGraw has been ghostwriting for more than twenty years. Last year she sent a writing sample for a book proposal job, and several AI detectors flagged it. She told NPR she hadn't used AI to generate any of it. The client said the detector results weren't the deciding factor and pointed to experience and fit instead. McGraw was paid for the sample, but she believes the flag cost her the full project. Her response to the whole episode is the part that sticks: she now steers around certain words and sentence structures, and she has stopped using some commas and em dashes. NPR published her story on October 31, 2026. Nothing about it is resolved. There is no rule that says a client can't run your draft through a detector, no standard for what a score means, and, as the article itself points out, no formal guidance for a writer who gets flagged and has to respond. We haven't tested any AI detectors ourselves, so this piece makes no claims about which ...
.png)