Grammarly's "Expert Review" AI feature provides writing suggestions "inspired by" various subject matter experts, including many journalists and public figures. The Verge reports that numerous experts, including its own staff, were included without their permission or knowledge, and some associated job titles are outdated.
The feature's implementation has significant issues: it frequently crashes, and its source links often lead to spammy websites, archived pages, or completely unrelated content, misrepresenting the basis for its advice. Grammarly's parent company defends the practice by stating the experts' published works are publicly available, but the presentation in tools like Google Docs can misleadingly simulate direct feedback from the named individuals.
The main topics covered are the unauthorized use of individuals' identities in an AI feature, the technical flaws and inaccuracies in the feature's execution, and the ethical concerns regarding permission and misleading presentation.
Grammarly’s “expert review” feature offers to give users writing advice “inspired by” subject matter experts, including recently-deceased professors, as Wired reported on Wednesday. When I tried the feature out myself, I found some experts that came as a surprise for a different reason — one of them was my boss.
Grammarly is using our identities without permission
‘Expert Review’ AI agents make suggestions supposedly inspired by subject matter experts, including several staff members here at The Verge.
‘Expert Review’ AI agents make suggestions supposedly inspired by subject matter experts, including several staff members here at The Verge.
The AI-generated feedback included comments that appeared to be from The Verge’s editor-in-chief, Nilay Patel, as well as editor-at-large David Pierce and senior editors Sean Hollister and Tom Warren, none of whom gave Grammarly permission to include them in the “expert reviews.”
The feature, which launched in August, claims to help you “sharpen your message through the lens of industry-relevant perspectives.” When users select the “expert review” button in the Grammarly sidebar, it analyzes their writing and surfaces AI-generated suggestions “inspired by” related experts. Those “industry-relevant perspectives” include the likes of Stephen King, Neil deGrasse Tyson, and Carl Sagan, among many others.
The Verge found numerous other tech journalists named in the feature, as well, including former Verge editors Casey Newton and Joanna Stern, former Verge writer Monica Chin, Wired’s Lauren Goode, Bloomberg’s Mark Gurman and Jason Schreier, the New York Times’ Kashmir Hill, The Atlantic’s Kaitlyn Tiffany, PC Gamer’s Wes Fenlon, Gizmodo’s Raymond Wong, Digital Foundry founder Richard Leadbetter, Tom’s Guide editor-in-chief Mark Spoonauer, former Rock Paper Shotgun editor-in-chief Katharine Castle, and former IGN news director Kat Bailey. The descriptions for some experts contain inaccuracies, such as outdated job titles, which could have been accurately updated had Superhuman asked those people for permission to reference their work.
In a statement to The Verge, Alex Gay, vice president of product and corporate marketing at Grammarly parent company Superhuman, commented: “The Expert Review agent doesn’t claim endorsement or direct participation from those experts; it provides suggestions inspired by works of experts and points users toward influential voices whose scholarship they can then explore more deeply.”
When asked if Superhuman considered notifying the people named in its AI feature, or requesting their permission, Gay said, “The experts in Expert Review appear because their published works are publicly available and widely cited.”
However, the experts’ work proved difficult to “explore more deeply.” The feature crashed frequently and its “sources” linked to spammy copies of legit websites, or other archived copies that aren’t the actual source page.
“The experts in Expert Review appear because their published works are publicly available and widely cited.”
Some sources even went to completely unrelated links that weren’t written by the person whose work they were supposedly an example of, potentially indicating that the suggestions Grammarly’s AI offers with one person’s name may be based on a different person’s work. This is only apparent if users click “see more” to expand suggestions, then click the “source” button at the end of the suggestion.
Additionally, the way the suggestions are presented could be misleading. In Google Docs, the suggestions look similar to comments from real users, seemingly simulating the experience of receiving edits from whichever expert the AI is imitating. One suggestion from Grammarly’s AI “inspired by” Verge senior editor Sean Hollister was about adding a parenthetical with context that was already included elsewhere. The only problem is that I’ve actually been edited by the real Sean Hollister, who prefers avoiding repetitive or unnecessary explanations while using straightforward wording and organization.
If I’d taken that advice and run it by him, the real Sean probably would have removed the parenthetical Grammarly suggested. An AI might be able to ingest vast amounts of someone’s writing and learn to mimic it, sure, but the same strategy cannot teach an AI how to edit the way that person would, based only on the writing they’ve published, even if you give the bot a check mark logo and call it an “expert.”
Meridiano
