As the lines between human and artificial intelligence continue to blur, a new feature from Grammarly has raised eyebrows in the academic and literary communities. The company's AI-powered writing tool now offers "expert" reviews from a roster of celebrated authors and scholars, both living and deceased.
This development reflects a defining technological shift in the writing assistance space, where AI-driven solutions are increasingly being positioned as substitutes for human expertise. According to CEO Shishir Mehrotra, the rebranded Superhuman suite of products aims to make technology feel "ordinary," but the implications of this "expert review" feature are anything but.
Inside the Platform
The Grammarly platform has evolved significantly over the past few years, incorporating a range of generative AI features designed to assist with writing tasks. The "expert review" option, in particular, has sparked controversy due to its use of AI agents modeled after prominent figures, including Stephen King, Neil deGrasse Tyson, and the late William Zinsser and Carl Sagan.
As Jen Dakin, senior communications manager at Superhuman, explains, the Expert Review agent uses the company's underlying large language model (LLM) to surface expert content and provide suggestions inspired by the works of these individuals. However, the disclaimer accompanying this feature explicitly states that the references to experts are for informational purposes only and do not indicate any affiliation or endorsement.
The Ethics of AI-Generated Expertise
The use of AI to simulate expert reviews has been met with skepticism and criticism from scholars, who argue that this approach validates the profound mistrust many have for AI and its applications in the humanities. C.E. Aubin, a historian and postdoctoral fellow at Yale University, contends that these "expert" reviews are, in fact, not expert reviews at all, as they do not involve actual human expertise.
Vanessa Heggie, an associate professor of the history of science and medicine at the University of Birmingham, has also spoken out against the feature, accusing Superhuman of creating "little LLMs" based on the scraped work of living and dead scholars, and trading on their names and reputations. The issue of reanimating the dead in this way has been particularly galling for many in the academic community.
Effectiveness and Implications
Beyond the ethical concerns, there are questions about the effectiveness of these AI-powered review tools. Grammarly's plagiarism detector, for example, failed to catch a direct quotation from The Simpsons, highlighting the limitations of these systems. As teachers and professors struggle to address the proliferation of AI-written essays, the introduction of "expert" reviews may inadvertently contribute to the perception that students are only double-checking their work, rather than violating academic codes of conduct.
The long-term implications of this technology are far-reaching, with potential consequences for the role of human instructors and the value placed on original thought and scholarship. As the boundaries between human and artificial intelligence continue to blur, it is essential to critically evaluate the impact of these developments on our educational systems and cultural institutions.
