Superhuman, the tech firm behind the popular writing assistant Grammarly, has abruptly disabled a controversial generative artificial intelligence (AI) feature that mimicked the prose styles of famous authors and journalists without their consent.
The move follows a wave of public backlash and a federal class-action lawsuit seeking damages of more than $5 million.
The Expert Review function, launched as part of a premium suite in August 2025, promised subscribers “inspired” feedback from digital personas modeled after figures such as Stephen King, Carl Sagan, and Neil deGrasse Tyson. However, the tool drew immediate fire from the very writers it sought to emulate, who argued their reputations were being “misappropriated” for corporate profit.
The legal challenge is led by Julia Angwin, a Pulitzer Prize-winning investigative journalist known for her work on tech privacy. The lawsuit, filed Wednesday in the Southern District of New York, alleges that Grammarly and Superhuman violated privacy and publicity rights by using the names and identities of hundreds of professionals to sell a $144-per-year subscription.
“I have worked for decades honing my skills as a writer and editor, and I am distressed to discover that a tech company is selling an imposter version of my hard-earned expertise,” Angwin said in a statement.
Beyond the lack of consent, Angwin criticized the quality of the AI’s output, describing advice from her “digital doppelgänger” as “scattershot” and counterproductive. In some instances, the AI suggested revisions that made text more complex and harder to understand — the antithesis of her professional style.
Superhuman CEO Shishir Mehrotra issued an apology on LinkedIn, acknowledging that the tool had “misrepresented” expert voices. While the company initially attempted to quell the unrest by offering an opt-out system, critics, including gaming journalist Wes Fenlon, blasted the move as “laughably inadequate.”
“We hear the feedback and recognize we fell short on this,” Mehrotra said, though he simultaneously defended the company’s legal position, calling the claims in the lawsuit “without merit.”
The feature reportedly utilized third-party Large Language Models (LLMs) trained on publicly available data to simulate the “voices” of experts. While a disclaimer noted that the writers had not participated in the tool’s development, legal experts argue that New York and California laws strictly prohibit the commercial use of a person’s likeness without permission.
The controversy highlights an escalating tension between the generative AI industry and the creative community. As tech firms race to monetize persona-based AI, writers and researchers — including AI ethicist Timnit Gebru, who was also included in the tool — are increasingly pushing back against the unauthorized commodification of their intellectual identities.
Superhuman says it is now “reimagining” the feature to ensure experts have control over their representation. For now, however, the Expert Review avatars of some of the world’s most prominent voices remain offline.
