Grammarly’s AI Expert Review Feature and the Misappropriation of Identities
Grammarly’s AI-powered writing assistant, Expert Review, generated editing suggestions and presented them under the names of real journalists, authors, and academics without their consent. The feature used a large language model to simulate the writing style and voice of individuals such as Stephen King and Carl Sagan, attributing feedback to them as if they had personally reviewed the user’s work.
The controversy arose when users discovered that the AI-generated feedback was falsely attributed to high-profile figures who had never authorized its use. The feature’s design created the illusion that these individuals were endorsing or providing personalized advice, which violated their intellectual property and reputational rights. The misappropriation extended beyond mere attribution—it also involved the commercial exploitation of their names and reputations for Grammarly’s profit.
The incident aligns with broader concerns about AI systems exploiting human identities without explicit permission. Such practices undermine trust in AI tools and raise ethical questions about consent, transparency, and the boundaries of AI-generated content. The case also highlights the risks of AI systems being deployed in ways that mislead users into believing they are interacting with real, human experts rather than automated processes.
Legal and Regulatory Consequences
The misuse of identities in Grammarly’s Expert Review feature triggered legal action, including a federal class action lawsuit filed by journalist Julia Angwin. The lawsuit alleged that Grammarly violated state privacy laws by using the names and likenesses of journalists, authors, and academics for commercial purposes without prior consent. Angwin’s legal claim argued that the feature constituted a form of identity theft, as it falsely represented that these individuals had provided personalized feedback on users’ writing.
The lawsuit further asserted that Grammarly’s actions harmed the reputations of the impersonated individuals by associating their names with AI-generated content that they had no control over. This misrepresentation could lead to confusion among users, who might assume that the feedback reflected the genuine opinions of the named experts.
Grammarly’s response to the backlash was swift: the company removed the Expert Review feature entirely shortly after the incident gained public attention. While the removal mitigated immediate harm, it did not address the broader legal or ethical implications of the feature’s design. The incident serves as a case study in how AI systems can inadvertently—or intentionally—cross ethical and legal lines when deployed without proper safeguards.
Harm to Epistemic Integrity and Reputational Damage
By presenting AI-generated feedback as if it came from real experts, the feature undermined the credibility of both the individuals whose names were used and the AI tool itself. Users who relied on the feedback may have been misled into believing they were receiving insights from trusted professionals, rather than algorithmically generated suggestions.
The reputational damage extended to the impersonated individuals, who had no involvement in the creation or endorsement of the AI-generated content. For example, authors like Stephen King and scientists like Carl Sagan have built careers on their intellectual contributions, and their names carry significant weight in their respective fields. Associating their identities with an AI tool that provided generic feedback without their consent could have diluted the perceived value of their work and created unnecessary confusion among their audiences.
Beyond the direct harm to the impersonated individuals, the incident also affected Grammarly’s own reputation. The case highlights the importance of obtaining explicit consent from individuals before their identities are used in AI-driven applications, particularly when those applications involve commercial or public-facing interactions.
Technical and Ethical Failures in AI System Design
The Expert Review feature’s design flaws reveal broader challenges in the development of AI systems that interact with human identities. The feature’s lack of transparency—failing to inform users that the feedback was AI-generated rather than human-provided—contributed to the deception.
If an AI tool is designed to mimic the style or voice of a real person, it must either obtain explicit consent from that individual or clearly disclose that the output is synthetic. Grammarly’s failure to implement either safeguard resulted in a violation of ethical AI principles, particularly those related to fairness, accountability, and transparency.
Instead, the company proceeded with a feature that relied on the commercial exploitation of others’ reputations, ultimately leading to legal and reputational consequences.
Broader Implications for AI and Writing Assistants
For AI developers, the Grammarly case serves as a cautionary tale about the unintended consequences of designing systems that blur the line between human and machine-generated output. Moving forward, companies must prioritize transparency by clearly labeling AI-generated content and obtaining explicit consent when using real individuals’ identities. Failure to do so risks not only legal repercussions but also long-term damage to user trust and public perception of AI technologies.
The incident also prompts a broader discussion about the role of AI in creative and professional fields. Writing assistants and other AI tools that interact with human-generated content must adhere to ethical standards that protect both users and the individuals whose identities or work may be indirectly affected. As AI continues to evolve, the industry must establish clearer boundaries to prevent similar misuses of identity and reputation in the future.
Sources
Incident summary from the AI Incident Database, used under CC BY-SA 4.0.