Grammarly’s Expert Review experiment has moved from backlash to the courtroom.
A new class action lawsuit alleges that the company used the names of real writers and experts to present AI-generated feedback, giving the impression that those professionals were personally reviewing users’ work. According to WIRED, the case pushes the controversy beyond product criticism and into a legal battle over consent, reputation, and the commercial use of identity in AI tools.
Inside the complaint
The case puts a name at the front of the fight. Julia Angwin, an investigative journalist and founder of Proof News, is the named plaintiff in a proposed class action that accuses AI writing tool Grammarly and its parent company, Superhuman, of using the identities of hundreds of writers and other public-facing professionals as part of a commercial AI feature.
The complaint argues the companies did more than draw on public writing. It says they packaged real names, reputations, and perceived expertise into a commercial product, then generated feedback that could be mistaken for guidance from people who never agreed to participate.
In the filing’s telling, that is the main problem: not just borrowed ideas, but borrowed identity.
The lawsuit did not come out of nowhere
Days before the case landed, Expert Review was already drawing heat for presenting AI-generated feedback as coming from real people. That criticism had already put the feature on shaky ground and sharpened attention on how Grammarly was using recognizable identities inside the product.
The lawsuit turns that public discomfort into something more concrete. What had been a growing controversy over product design is now being framed in legal terms, with the complaint pressing the question of whether that use of identity crossed a line.
Expert Review is taken offline
Superhuman had already begun pulling the plug. In a LinkedIn post, CEO Shishir Mehrotra said the company was disabling Expert Review after criticism that the feature misrepresented experts’ voices, conceding that it “missed the mark” even as he defended the original ambition behind it.
The retreat did not come with a full retreat from the idea itself. Mehrotra said the company now wants to rethink the feature around a model where experts have more control over how they are represented, a sign that the rollout may be over for now, but the larger effort to build AI tools shaped around recognizable expertise is not.
A federal injunction against Perplexity’s Comet browser gives Amazon an early edge in a closely watched AI access fight.