Remember when Grammarly was just a glorified spellchecker that nagged you about passive voice? Those days are long gone. The company, now awkwardly rebranded as “Superhuman” (yes, really), has been cramming generative AI features into the platform for years. But their latest stunt crosses a line that should make anyone with a shred of intellectual property sense uneasy.
Grammarly now offers an “expert review” feature that lets you solicit feedback from AI agents modeled on real writers and scholars—living and dead. Want Stephen King to critique your thriller draft? Neil deGrasse Tyson to weigh in on your science paper? Carl Sagan to comment on your astronomy essay? The software will happily generate a simulation of their advice, trained on their published work, without their permission or involvement.
The company’s disclaimer is buried in fine print: “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.” In other words, they’re selling the illusion of authority while admitting it’s entirely fabricated.
Vanessa Heggie, a historian at the University of Birmingham, called this out on LinkedIn after discovering the system offered feedback from an AI version of David Abulafia—a medieval historian who died in January. Her verdict: “Obscene.” I’m inclined to agree.
Let’s be clear about what’s happening here. Grammarly is scraping the published works of living authors and academics, plus the estates of dead ones, to build LLM-based personas without consent. The legal landscape for this is a minefield—there are already multiple copyright lawsuits targeting similar practices. But more importantly, it’s just gross. These are people’s life’s work, their voice, their reputation. Reducing that to a drop-down menu of critique bots feels like something out of a Black Mirror episode.
C.E. Aubin, a historian at Yale, put it well on Bluesky: “These are not expert reviews, because there are no ‘experts’ involved in producing them.” She’s right. The feature doesn’t just misrepresent the people it mimics—it actively undermines the value of actual human expertise. Why pay a real editor or consult a living scholar when you can get a cheap simulacrum? The message is clear: the person behind the work is irrelevant, only the output matters.
There’s also the question of whether this is even useful. I tested the feature (well, a colleague did) and the feedback was laughably generic. The AI version of Virginia Tufte, who died in 2020, offered the profound insight: “Replace repetition with vivid, varied sentence patterns.” Groundbreaking stuff. You could get that from any chatbot without exploiting someone’s legacy.
Grammarly’s senior communications manager, Jen Dakin, defended the feature by saying it “doesn’t claim endorsement or direct participation from those experts” and is meant to “point users toward influential voices whose scholarship they can then explore more deeply.” That’s a generous framing. A less charitable reading is that they’re using dead people’s names to sell subscriptions, and living people’s work without compensation.
Neither Stephen King nor Neil deGrasse Tyson responded to requests for comment. I can’t imagine why.
The broader trend here is troubling. We’re watching companies rush to monetize every scrap of human creativity, treating decades of scholarship and artistry as free training data. The academics and authors being mined here aren’t tech CEOs—they’re people who dedicated their lives to thinking deeply and writing carefully. Reducing that to a parlor trick for students who don’t want to visit office hours is insulting.
I’m not a Luddite. AI has real uses in writing: catching errors, suggesting alternatives, helping with structure. But there’s a difference between a tool that assists and one that impersonates. Grammarly’s expert review feature is firmly in the latter camp, and it leaves a bad taste.
If you’re going to get writing advice, get it from a human—one who consented to give it. Or at least from an AI that doesn’t pretend to be someone else.
Comments (0)
Login Log in to comment.
Be the first to comment!