Grammarly Eliminates AI Expert Review Feature That Imitated Writers Following Criticism

Grammarly has recently taken the decisive step to disable a controversial AI feature known as Expert Review, which was designed to mimic the writing styles of well-known authors and academics. This move comes amid mounting legal pressure, as a multimillion-dollar lawsuit has been filed against the company by individuals whose identities were leveraged without their consent.

The Expert Review feature employed generative AI technology to deliver feedback that purports to be inspired by the prose of notable figures such as novelist Stephen King, astrophysicist Neil deGrasse Tyson, and the late Carl Sagan, who was a revered scientist and science communicator. This feature aimed to provide personalized feedback to users by tapping into the expertise of these distinguished writers.

A class-action lawsuit has been initiated in the Southern District of New York against Superhuman, the parent company of Grammarly. The plaintiffs contend that the unauthorized use of their names for commercial purposes is illegal. They further assert that the damages owed to the class of complainants exceed $5 million (£3.7 million), indicating the serious nature of the allegations.

Since the controversy surrounding Grammarly’s feature has gained traction, several writers have come forward expressing their discontent about being included without their permission.

Tech journalist Casey Newton, who found himself featured in the software, voiced his concerns by stating, “Grammarly curated a list of real people, gave its models free rein to hallucinate plausible-sounding advice on their behalf, and put it all behind a subscription. That’s a deliberate choice to monetize the identities of real people without involving them, and it sucks,” underscoring the ethical issues at stake in the situation.

Vanessa Heggie, an associate professor at the University of Birmingham, took to LinkedIn to express her outrage. She highlighted that her late colleague, David Abulafia, who passed away in January, was also included in the feature. She described the use of such identities without consent as “obscene,” illustrating the broader sentiment of violated trust felt by many academics.

Julia Angwin, a notable investigative journalist who was also represented in the application, has stepped up as the lead plaintiff in the lawsuit. In an interview with the BBC, she remarked, “I had thought of deepfakes as something that happens to celebrities, mostly around images. Editing is a skill… it’s my livelihood, but it’s not something I’ve ever thought about anyone trying to steal from me before. I didn’t even think it was steal-able.” Angwin’s comments reveal a growing awareness of the complexities surrounding AI technology and identity.

Her attorney, Peter Romer-Friedman, noted that the case has rapidly garnered attention from the writing community. “We’ve heard from over 40 people in the last 24 hours since we filed the suit,” he stated, suggesting that this issue resonates profoundly with creators concerned about their intellectual property.

Initially launched in 2009 as a spelling and grammar correction tool, Grammarly expanded its offerings last year to include various AI-driven features, one of which was Expert Review. The company claimed that the Expert Review agent aimed to provide subject-matter expertise and tailored feedback to help users enhance their writing to meet rigorous academic or professional standards. They previously stated their commitment to empowering writers through assistance and support.

In the wake of the backlash, Superhuman’s Chief Executive, Shishir Mehrotra, issued an apology via LinkedIn. He acknowledged the valid criticisms received from experts who felt that the AI agent misrepresented their voices. “We hear the feedback and recognize we fell short on this. I want to apologize and acknowledge that we’ll rethink our approach going forward,” he mentioned, indicating the company’s willingness to reassess their methods.

In light of the ongoing lawsuit, Mehrotra further clarified that the discontinuation of the Expert Review feature was already in place for redesign purposes before the filing of the claim. He noted that the feature had limited usage during its brief existence. “We are sorry, and we will rethink our approach going forward,” he reiterated, emphasizing the company’s intent to make amends.

However, he also maintained that the legal claims against the company were “without merit” and promised that Superhuman would vigorously defend itself against these allegations, illustrating the depth of the conflict and the stakes involved.