Asking the Dead to Review Your Essay

David Abulafia died in January. By March, Grammarly had him reviewing your prose.

This is not a ghost story, though it has all the ingredients of one. The grammar and writing tool — used by an estimated 30 million people — launched a feature called "Expert Review," which offered users personalised writing feedback delivered through AI personas of celebrated writers and academics. Stephen King would opine on your sentence rhythm. Carl Sagan would weigh in on your clarity. And David Abulafia, the distinguished historian of the Mediterranean and Professor Emeritus at Cambridge, would assess your academic writing — two months after he had died.

The University of Birmingham's Vanessa Heggie, who studies the history of medicine and was herself among the impersonated experts, described Abulafia's inclusion as "obscene." Hard to argue. The communications director at Grammarly, for her part, clarified that the feature generated suggestions "inspired by works of experts," which is the kind of explanation that manages to be simultaneously accurate and completely beside the point.

The literary world has spent the better part of a decade arguing about what AI means for writers — for copyright, for voice, for the economic realities of the profession. These arguments tend to be large and abstract: training data, intellectual property, the future of creativity. The Grammarly affair is smaller and more specific, and perhaps for that reason more revealing. This was not a question of scale. It was a question of consent.

There is something particularly clarifying about the case of a dead man. Abulafia cannot consent. His estate has not been asked. His reputation — earned over decades of scholarship on the medieval Mediterranean, on commerce and conquest, on the interconnected world of the preindustrial sea — has been annexed to sell subscriptions. "Inspired by" does a lot of heavy lifting here. Updike was inspired by the New England Protestant tradition; that did not make him its spokesman.

Grammarly pulled the feature after the backlash. The company did not apologise, exactly — it explained. There is a difference, though tech companies tend to find it blurry.

The practical question now is who remains on guard. Grammarly will iterate; the next version of this feature will be subtler, the disclaimers smaller, the language more carefully calibrated. Stephen King, at least, can complain. The estate of Carl Sagan might have something to say. The academics and editors with smaller platforms — also listed as "experts," whose names have not made headlines — will have a harder time. They always do.

What strikes you, stepping back, is how routine this has all become: the launch, the outcry, the quiet withdrawal. We are no longer shocked by any of it. That, if nothing else, is what Grammarly should be worried about.