Writing Tool Removes AI Personas After Legal Challenge from Authors

Grammarly has disabled an AI feature that mimicked the writing styles of prominent authors and scientists without their consent, following a significant legal challenge from the writers whose identities were used. The Expert Review function, which offered writing feedback “inspired by” the personas of figures including Stephen King and Carl Sagan, was taken down this week by Superhuman, the tech firm that operates Grammarly. The move came after a multi-million dollar lawsuit was filed in the Southern District of New York by investigative journalist Julia Angwin and other writers who discovered their names and professional reputations being marketed as commercial AI personas. Superhuman’s chief executive acknowledged the tool had “misrepresented” the voices of experts|expert voices, apologizing for the controversial feature.

The Function That Sparked Public Backlash

The Expert Review function was a substantial shift from Grammarly’s traditional editing support approach. Rather than providing generic feedback, the tool permitted users to obtain editing suggestions “inspired by” the unique styles of celebrated writers and academics. Users could pick from personas including acclaimed novelist Stephen King and renowned scientist Carl Sagan, among numerous other public figures. The feature offered to offer customized writing advice viewed through the lens of these esteemed experts, ostensibly enabling users refine their work by drawing inspiration from the best in their respective fields.

What Grammarly marketed as an innovative learning platform quickly revealed itself as a concerning unauthorized use of identity and intellectual property. The company had never requested permission from any of the writers whose personas were being reproduced for profit. Investigative journalist Julia Angwin, who became the lead plaintiff in the class action suit, stated her surprise at discovering her professional reputation being offered as a product feature. She described the situation as fundamentally different from traditional deepfakes, emphasizing that editing expertise represents her livelihood and that she had never imagined her professional skills could be stolen and packaged this way.

• AI personas replicated hundreds of author voices without authorization or financial compensation
• Feature delivered recommendations drawn from renowned writers and researchers
• Users could pick different expert personas for revision feedback
• Tool was built into Grammarly’s paid subscription service offerings

Court Action and Market Reaction

The legal action against Superhuman and Grammarly represents a significant moment in the broader debate over artificial intelligence ethics and IP protections. Led by journalist investigator Julia Angwin, the class-action lawsuit submitted to the Southern District of New York alleges that the company illegally appropriated the identities of numerous writers to create revenue from its subscription-based service. The filing maintains that using names and professional reputations for commercial purposes absent explicit authorization breaches existing legal protections defending individuals from unlawful commercial use of their identity and likeness.

The reaction to the lawsuit has been rapid and significant. Within 24 hours of submission, Angwin’s legal team reported hearing from over 40 prospective claimants eager to join the action, illustrating broad apprehension among affected writers. The case pursues damages surpassing $5 million, though industry analysts suggest the actual figure could be considerably higher once the court determines compensation based on the company’s earnings from the controversial tool. Superhuman’s rapid decision to deactivate the Expert Review function suggests the company understood the reputational and legal risks created by continuing the feature.

The Lawsuit Details

The lawsuit specifically contends that Grammarly and Superhuman infringed upon core protections of identity privacy by attributing editorial advice to writers who never provided such direction. The lawsuit filing underscores that the organization profited from these personas through its paid membership structure, creating profits via the improper use of hundreds of individuals’ names and career standing. Attorneys contend this amounts to a “blatant violation of the law,” referencing established court rulings safeguarding people from commercial misappropriation of their identities without consent.

Julia Angwin’s personal frustration with the feature went further than the unlawful infringements to the quality of the AI’s output. She described the proposed edits assigned to her as a “slopperganger”—a term describing substandard machine-created text—pointing out that the edits were rendering sentences more problematic rather than enhancing them. This added dimension to the case underscores not only the regulatory violation but also the harm to reputation of having one’s identity connected to substandard professional work, intensifying the damage of improper use of one’s name.

• Damages sought exceed $5 million with actual figure based on company earnings
• Over 40 additional plaintiffs contacted legal team within the initial day
• Alleges unlawful commercial use of identities without consent or compensation

Trust Deficiencies and Reliability Problems

Beyond the legal violations, the Expert Review function prompted significant concerns about the dependability and precision of algorithmically-produced editorial guidance. Users relying on suggestions attributed to established writers and academics had no way of knowing whether they were receiving genuine guidance or computer-generated imitations of professional expertise. This loss of confidence goes further than individual plaintiffs to the wider writing sector, where readers and students might have legitimately believed they were receiving instruction from established authorities. The feature’s removal underscores a critical gap between what artificial intelligence can technically achieve and what it should ethically be permitted to do.

The reputational harm inflicted on impersonated writers turned out to be particularly insidious because it connected their identities with inferior quality work. Angwin’s experience exemplified this problem—her professional standing was being presented as a premium service while simultaneously delivering substandard editorial support. This pairing of unauthorized use and inferior standards caused dual harm: lack of control over her name coupled with connection to substandard results that conflicted with her professional standards. For writers whose reputation depends on the caliber of their output, such unauthorized use poses an existential threat to their professional standing and market position.

The Issue with AI Mimicry

The central weakness in Grammarly’s approach lay in attempting to replicate the subtle discernment and expertise of experienced authors through automated systems. Skilled revision requires context-specific knowledge, stylistic sensitivity, and decades of refined practice—elements that cannot genuinely be replicated by examining text examples and generating responses in a similar voice. Angwin’s point that the AI-generated edits made sentences more complex rather than enhancing quality exposed the essential inadequacy of the mimicry. The platform could reproduce obvious stylistic features but lacked the substantive grasp required for offering genuinely valuable editorial guidance, ultimately compromising both the genuineness of the voices and the practical value of the offering itself.

Corporate Response and Next Steps

Superhuman’s CEO Shishir Mehrotra recognized the misstep in public, issuing an apology on LinkedIn in which he admitted that the Expert Review function had “misrepresented” the voices of the experts being impersonated. The company’s quick move to disable the feature this week suggests an effort to mitigate further legal and reputational damage. However, the deactivation came only following the lawsuit was filed and substantial public criticism emerged, prompting concerns about whether the company would have taken action without outside pressure. Mehrotra’s statement, while apologetic in tone, did not tackle the broader question of how such a feature was greenlit and launched in the first place, nor did it detail specific steps to prevent similar incidents in the future.

The road ahead for Grammarly remains uncertain as the legal case continues. Beyond the urgent legal dispute, the company confronts the challenge of rebuilding trust with content creators and users who may now scrutinize the ethical standards governing its AI creation. The discontinuation of the Expert Review function constitutes a reactive rather than proactive stance, implying the company is addressing legal pressure rather than demonstrating genuine commitment to responsible AI development. Moving forward, Grammarly will likely need to establish stricter permission frameworks and oversight mechanisms for any capabilities that employ the utilization of real people’s names or likenesses. The company’s handling of this crisis may set precedent for how other AI firms address the utilization of celebrities in their generation systems.

The speed at which Grammarly deactivated the feature points to the company acknowledged the legal, reputational, and business stakes involved. However, the lack of proactive measures before the lawsuit demonstrates that review procedures failed to catch the ethical breaches. As the case moves forward, the company could encounter increased oversight concerning how numerous content creators were impacted and whether damages will be provided above what the lawsuit calls for. The case is likely to influence how other AI companies address the inclusion of real identities in their products going forward.