Grammarly’s New ‘Expert Review’ Lets You Get AI Feedback From Dead People (Without Their Permission)

Remember being a teacher’s pet? Or that professor whose margin notes actually made your writing better? Grammarly thinks you want that back—except now the feedback comes from a dead person’s AI ghost.

Grammarly, once just a glorified spellchecker, has been piling on generative AI features for years. In October, CEO Shishir Mehrotra announced the whole company rebranded as Superhuman, though the writing tool still goes by Grammarly. Mehrotra’s press release had the usual corporate fluff: “When technology works everywhere, it starts to feel ordinary.” Sure, Jan.

The expanded platform now offers an AI chatbot that answers questions mid-draft, a paraphraser, a “humanizer” that adjusts tone, an AI grader that predicts college coursework scores, and even tools to flag phrases that sound too much like an LLM. (Because nothing says authenticity like using AI to hide your AI use.)

But the most insidious addition is the “expert review” option. Instead of generic LLM feedback, you get to pick from a list of real academics and authors—living and dead—whose simulated voices will critique your text. Stephen King, Neil deGrasse Tyson, William Zinsser, Carl Sagan. The fine print admits these people have nothing to do with it: “References to experts in this product are for informational purposes only and do not indicate any affiliation with Grammarly or endorsement by those individuals or entities.”

Neither King nor Tyson responded to requests for comment. I’m shocked.

Superhuman’s senior comms manager Jen Dakin tried to spin it: “The Expert Review agent examines the writing… and leverages our underlying LLM to surface expert content that can help the document’s author shape their work.” The suggested experts “depend on the substance of the writing being evaluated.” It doesn’t claim endorsement, just provides suggestions “inspired by works of experts.”

Vanessa Heggie, an associate professor at the University of Birmingham, called this out on LinkedIn as “creating little LLMs” based on “scraped work” of the living and dead, trading on “their names and reputations.” She posted a screenshot showing an AI agent modeled on David Abulafia, a historian who died in January. Her verdict: “Obscene.”

WIRED tested the tool and got recommendations from bots based on Steven Pinker, Gary Marcus, William Strunk Jr., Pierre Bourdieu, Margaret Mitchell, and Virginia Tufte—all dead except Pinker and Marcus. The guidance from Tufte’s AI: “Replace repetition with vivid, varied sentence patterns.” Riveting.

C.E. Aubin, a Yale historian, told WIRED this “validates the profound mistrust so many scholars in the humanities have for AI and its seemingly constant use in fundamentally unethical ways.” She added: “These are not expert reviews, because there are no ‘experts’ involved in producing them.” Aubin called the elimination of personhood “awful” enough on its own, apart from “the issue of ‘reanimating’ the dead so cynically.”

Beyond the ethics, there’s the practical question: does this even work well? Grammarly’s plagiarism detector couldn’t catch a direct quote in WIRED’s test. Not exactly confidence-inspiring.

The whole thing feels like a tech company running ahead of consent, legality, and basic decency. The copyright lawsuits are piling up, and this feature just adds fuel. I’m not sure who asked for AI Stephen King to critique their marketing brief, but here we are.