Grammarly’s AI Used Real Authors’ Names Without Consent

Grammarly recently rolled out an "Expert Review" feature that aims to make AI writing suggestions feel more authoritative by attributing them to well-known writers. Journalists at multiple outlets discovered the tool using their real names — including some who never agreed to lend their bylines to an AI product.

The revelations began circulating after writers noticed AI comments and guidance labeled with recognizable author names in draft documents. Reports show several journalists and authors found their names used to add credibility to advice generated by Grammarly’s models, without prior consent or notification.

Rather than removing the feature or issuing an apology, Grammarly has responded by giving affected writers an option to opt out. That means users can request their names be excluded going forward, but the company is not yet retracting the practice or fully acknowledging the underlying problem.

Critics say the approach raises ethical questions: attributing AI output to real people can mislead users about the source and authority behind a suggestion, and it can risk reputational harm for those whose names are used without consent. Supporters of labeled AI notes argue that attaching a recognizable voice can help users evaluate tone and style suggestions — but the difference here is the lack of permission.

Grammarly positions the Expert Review as a way to offer stylistic perspectives, yet the rollout has exposed a gap between product design and the expectations of people whose names and work are invoked. For now, the company’s opt-out route may ease immediate concerns for individuals, but the episode underscores broader debates about transparency and consent as AI systems adopt more humanlike attribution strategies.