Grammarly’s New AI Tool Provides Writing Insights from Famous Scholars

Grammarly’s latest AI feature, designed to provide writing feedback from the perspective of so-called “experts,” is facing scrutiny from academics who feel this tool seems to “resurrect” scholars for reviewing users’ work.

Known as Expert Review, this feature analyzes texts and generates feedback based on the viewpoints of specific scholars, journalists, and other authorities. Worryingly, many of the so-called experts the AI tool references are deceased—a characteristic that one medieval historian on BlueSky described as “morbid.”

Originally launched in 2009 as an AI-assisted writing aid, Grammarly rebranded to Superhuman last October to signify its evolution from a simple writing assistant to a suite of AI productivity tools, encompassing services for research, scheduling, email, and workflow automation.

The Expert Review feature was introduced last summer, allowing users of the Superhuman Go version to select an expert and receive AI-generated feedback aligned with that scholar’s field or published works through the Grammarly browser extension.

“Our Expert Review agent assesses the writing a user is working on, whether it’s a marketing brief or a student project about biodiversity, and utilizes our underlying LLM to highlight expert insights to help the author refine their work,” a Superhuman representative stated to Decrypt. “The recommended experts depend on the content of the writing being evaluated.”

The spokesperson clarified that the Expert Review agent does not claim endorsement or direct involvement from the experts, but instead provides “suggestions influenced by the works of those experts and guides users toward significant voices in their field for further exploration.”

“The experts mentioned in Expert Review are included because their works are publicly accessible and broadly cited,” they added.

While the feature intends to assist students and professionals in enhancing their writing skills, Vanessa Heggie, a professor of history at the University of Birmingham, raised concerns over whether the “reviewers” had given permission for their inclusion in the app.

“I don’t know where to start with this, but… Grammarly is now offering ‘expert review’ of your work by living and deceased academics,” Heggie wrote on LinkedIn. “Yes, even deceased ones—without anyone’s explicit permission, it is generating little LLMs based on their scraped work and using their names and reputations. Absurd.”

Brielle Harbin, a former associate professor of political science at the United States Naval Academy, labeled it “an odd and concerning evolution.”

“Options like this—especially made without context, consent, or substantial collaboration with educators—may deepen skepticism regarding AI tools in higher education,” she wrote on LinkedIn. “Ironically, initiatives intended to enhance adoption may inadvertently reinforce resistance. Trust and collaboration are crucial at this moment.”

Grammarly is not the only organization creating artificial intelligence programs designed to replicate real individuals.

In 2023, Meta introduced a lineup of chatbots for its Meta AI platform centered around celebrity personas, including Snoop Dogg, Tom Brady, Kendall Jenner, and Naomi Osaka. This year, Khan Academy also launched its AI tutor Khanmigo, enabling students to engage in simulated dialogues with historical figures such as British Prime Minister Winston Churchill and U.S. Civil War spy and Underground Railroad conductor Harriet Tubman.