Grammarly, the writing assistant recently rebranded under the parent company Superhuman, has sparked intense ethical backlash by launching an “Expert Review” feature that uses generative AI to simulate feedback from legendary authors and scholars without their consent. This strategic pivot, announced by CEO Shishir Mehrotra, transforms the tool from a simple grammar checker into an AI-driven “writing partner” capable of harvesting the intellectual legacies of figures like Stephen King, Carl Sagan, and the late William Zinsser.
The Evolution from Proofreader to AI ‘Superhuman’
The platform’s rebranding reflects a massive expansion of its generative capabilities. Beyond standard spellcheck, Grammarly now offers a sophisticated suite of AI utilities designed to automate every stage of the writing process. These include a “paraphraser” for stylistic adjustments, a “humanizer” that masks AI-generated patterns, and an automated grader that predicts academic performance for college-level coursework.
CEO Shishir Mehrotra defends the integration, suggesting that when technology becomes pervasive, it signals “something extraordinary” occurring beneath the surface. However, the most controversial addition is the “Expert Review” agent, which promises users feedback “inspired by” the world’s leading thinkers.
Unauthorized Impersonation and the Ethics of ‘Digital Reanimation’
The “Expert Review” tool presents users with a roster of virtual mentors, including living scientists like Neil deGrasse Tyson and deceased luminaries such as astronomer Carl Sagan or historian David Abulafia. Despite using these names to market the service, Grammarly includes a disclaimer stating that these experts have no affiliation with the product and have not endorsed its findings. The system leverages Large Language Models (LLMs) to analyze a user’s text and surface content that mimics the specific tone and scholarship of these figures.
Academic Backlash Against ‘Scraped’ Intellectual Property
Scholars and historians have reacted with vitriol to the feature. Vanessa Heggie, an associate professor at the University of Birmingham, publicly condemned the company for “creating little LLMs” based on the scraped work of both the living and the dead. Heggie characterized the use of David Abulafia’s persona—a medieval historian who passed away recently—as “obscene.”
Similarly, Yale University postdoctoral fellow C.E. Aubin argues that the system validates the deep-seated mistrust within the humanities regarding unethical AI deployment. Aubin contends that reducing complex scholarship to an algorithm eliminates personhood and insults the actual thinkers whose work is being “cynically reanimated” for commercial gain.
Technical Limitations and Plagiarism Risks
Independent testing reveals significant gaps in the tool’s effectiveness. In a review conducted by WIRED, Grammarly’s plagiarism detector failed to identify a direct, verbatim quote from a well-known television script, though it did flag certain phrases as “commonly generated by LLMs.” This creates a dangerous paradox for students: the platform provides tools to make writing sound more “human” and “expert-reviewed,” yet these very features may lead users to inadvertently violate academic integrity codes.
As educators struggle to manage a surge in AI-assisted essays, the “Expert Review” feature offers a shortcut that bypasses genuine intellectual engagement. By replacing human instructors with illusory AI mentors, the platform shifts the educational landscape toward a model where the “thinkers” are entirely removed from the equation, leaving only the algorithm to judge the output.
