Grammarly just shut down its "Expert Review" feature and now faces a class action lawsuit over it. The reason: the feature presented AI-generated editing suggestions as if they came from established authors and academics, using real names and credentials without those people's knowledge or permission.
That's not a small misstep. It's fabricating endorsements.
What the Feature Actually Did
Grammarly's Expert Review was marketed as a premium layer of human-quality feedback. Users saw editing suggestions tagged with names and bios of real writers and scholars, implying those people had personally reviewed the text. In reality, the suggestions were AI-generated. The named experts had no involvement, had not reviewed any documents, and in many cases had no idea their identities were being used.
The class action alleges this misled paying customers who believed they were getting genuine expert feedback, and violated the rights of the individuals whose names and likenesses were used without consent.
Grammarly pulled the feature on Wednesday, but the legal challenge moves forward.
A Trust Problem for AI Writing Tools
This hits at something broader than one feature. AI writing assistants occupy a unique position - users trust them with drafts of emails, reports, essays, and business documents. That trust depends on the tool being transparent about what it is and is not doing.
Presenting AI output under a human expert's name crosses a line that most users would consider obvious. It's the kind of decision that makes you wonder what internal review process approved it in the first place.
For Grammarly's 30+ million daily users, the practical impact of losing Expert Review is minimal since most rely on the core grammar and style checking. But the reputational damage could linger, especially as competing tools like ProWritingAid and Wordtune push their own AI editing features.
The lawsuit also sets a precedent worth watching. If AI companies can be held liable for attaching real people's names to AI-generated output, that changes the calculus for any product that blends human branding with machine output. Marketing pages full of AI-generated testimonials attributed to real users, chatbots that impersonate named advisors, content tools that fake bylines - all of those look riskier after this filing.
Grammarly has not commented publicly on the lawsuit beyond confirming the feature was removed.