Grammarly Uses Journalists' Identities Without Permission for AI Agents

The incident came to light when Verge staff members noticed their colleagues' names, biographical details, and professional expertise being leveraged by Grammarly's AI system to provide supposedly authoritative feedback on written content. These Expert Review agents were presented to users as knowledgeable professionals offering specialized insights, but they were actually AI constructs masquerading as real people with established careers and reputations.

Grammarly's Expert Review feature promises users access to subject matter expertise across various fields, from journalism and marketing to technical writing and academic research. The system analyzes user content and provides suggestions framed as coming from industry professionals with relevant experience. However, the revelation that these "experts" are unauthorized digital impersonations of real people fundamentally undermines the feature's credibility claims.

The Mechanics of Digital Identity Theft

The process appears straightforward yet deeply problematic: Grammarly's systems scraped publicly available information about professionals, including their work history, published articles, and areas of expertise. This data was then used to construct AI personas that could reference specific knowledge domains and provide contextually relevant writing advice. The AI agents weren't simply generating generic feedback—they were doing so while explicitly borrowing the professional authority and reputation of real individuals.

What makes this particularly concerning is the level of detail involved. These weren't generic "marketing expert" or "journalist" labels, but rather specific identities tied to real people with verifiable careers and public profiles. The system essentially created digital doppelgangers that could interact with users while trading on the professional accomplishments of people who had no knowledge of or consent to this use.

The practice extends beyond simple name usage into more sophisticated forms of identity appropriation. The AI agents incorporated writing styles, subject matter preferences, and professional perspectives that aligned with their real-world counterparts' public work. This level of personalization suggests a systematic effort to create convincing expert personas rather than a simple oversight or automated error.

Industry Implications and Broader Pattern

Grammarly's unauthorized use of professional identities reflects a growing trend among AI companies seeking to bridge the credibility gap between artificial and human-generated content. As users become more sophisticated in identifying AI-generated text, companies are increasingly looking for ways to add human authority to their automated systems. However, borrowing real people's identities without permission crosses ethical and potentially legal boundaries.

This incident parallels other recent controversies in the AI space, where companies have used copyrighted material, personal data, and professional work without proper authorization. The difference here lies in the direct impersonation aspect—rather than simply training on someone's work, Grammarly was actively presenting AI outputs as coming from specific, named individuals.

The implications extend beyond individual privacy concerns to questions about professional liability and reputation management. When an AI system makes suggestions or provides advice while using someone's identity, that person could potentially be held responsible for the quality or accuracy of that guidance. Journalists and subject matter experts build their reputations over years of careful work, and having their names attached to AI-generated content they cannot control poses significant professional risks.

The Authentication Challenge

For users of AI-powered creativity tools, this revelation highlights the difficulty of verifying the authenticity of claimed human involvement. Grammarly marketed its Expert Review feature as providing access to real professional expertise, but users had no way to verify whether the suggestions they received actually came from the claimed sources or were generated by AI systems using those people's identities.

The incident raises broader questions about how AI companies should handle the integration of human expertise into their systems. While there are legitimate ways to incorporate professional knowledge—through paid partnerships, licensed content, or explicit consent agreements—the unauthorized use of identities represents a shortcut that prioritizes user experience over individual rights and professional ethics.

As AI writing tools become increasingly sophisticated and widespread, the pressure to differentiate through claimed human expertise will likely intensify. Companies may feel compelled to find creative ways to add authenticity to their offerings, but the Grammarly incident demonstrates the risks of crossing ethical boundaries in pursuit of competitive advantage. The technology industry must grapple with establishing clear standards for consent and attribution as AI systems become more capable of mimicking human expertise and authority.

AUTHOR

Alexa Hill

Coffee lover, sushi whisperer. Alexa is a graphic designer turned user experience wizard who knows all the details of every image format out there.