Grammarly expert review feature borrows real names

Grammarly has been pitching a feature called “Expert Review” as a way to improve writing “from the perspective” of famous authors, public intellectuals and working journalists. In practice, it is a layer of AI-generated feedback that can appear to be voiced by real people — including reporters at outlets such as The Verge, Wired, Bloomberg and The New York Times — without any indication that those individuals participated or gave permission.

According to TechCrunch, the feature sits inside Grammarly’s main writing assistant and invites users to summon revision notes framed as if they come from “the world’s great writers and thinkers.” Wired and The Verge reported that the product presents suggestions as if they were authored by named figures, living and dead, even though Grammarly’s own documentation treats the names as mere “references.” A vice president at Grammarly’s parent company Superhuman told The Verge that the experts are mentioned because their published works are “publicly available and widely cited,” and Grammarly’s user guide adds a disclaimer that the references “do not indicate any affiliation … or endorsement.”

That legal language is doing a lot of work. The everyday user experience is not a footnote; it is the product. A sidebar labeled “Expert Review” is a promise of accountability: someone qualified has looked at your text and judged it. Many buyers — especially managers procuring tools for teams — want the feeling of quality control more than they want to pay for the costly part of quality control, which is time from a real editor who can be named, questioned and blamed.

This is the market gap AI is filling: not just drafting and rewriting, but simulating institutional sign-off. Once the “expert” is a UI element rather than a person, responsibility diffuses. The employee can say the tool recommended it; the manager can say the tool was purchased to reduce risk; the vendor can point to a disclaimer. Meanwhile the reputational capital being borrowed is real: a journalist’s byline and an author’s name are treated as props to make machine output feel like review.

The oddest detail in TechCrunch’s test is how the feature manufactures editorial mannerisms — “add ethical context like Casey Newton,” “leverage the anecdote for reader alignment like Kara Swisher” — without any relationship to the people invoked. It is not expertise so much as brand-flavoured style transfer.

Grammarly’s guide says the names are “for informational purposes only.” The button still says “Expert Review.”