If you ever cherished the idea of being a teacher's favorite or craved the feedback from a beloved college professor, a recent innovation may resonate with you. A prominent software company has designed a way to mimic critiques not only from contemporary bestselling authors and esteemed academics but also from those who passed away long ago—all without securing consent from anyone involved.
Initially designed to assist with grammar and spelling corrections, Grammarly has significantly expanded its capabilities in recent years, incorporating several generative AI features. In October, CEO Shishir Mehrotra revealed a company rebranding as “Superhuman” to showcase a new range of AI-driven products, yet the writing assistance tool itself retains the name Grammarly. As Mehrotra noted in his press release, “When technology works everywhere, it starts to feel ordinary, which usually means something extraordinary is happening beneath the surface.”
The updated Grammarly platform presents an AI solution to almost every conceivable writing requirement—many that you may not have even thought of. It features an AI chatbot for real-time questions while drafting, a “paraphraser” to suggest style alterations, a “humanizer” to revise text in a specific tone, an AI grader to predict potential scores for academic assignments, and tools to identify and refine phrases often associated with large language models. (Although you’re relying on AI for assistance, you probably want to avoid it sounding automated.)
One of the more controversial additions is the “expert review” option. Instead of generic feedback from an anonymous language model, this feature connects users with the names of real authors and academics who might offer insights on their writing. Importantly, these individuals have no actual involvement in the process. As stated in a disclaimer, “References to experts in this product are for informational purposes only and do not indicate any affiliation with Grammarly or endorsement by those individuals or entities.”
According to a support page, Grammarly users can request advice from virtual versions of both living writers and scholars like Stephen King and Neil deGrasse Tyson, alongside the deceased, such as editor William Zinsser and astronomer Carl Sagan. It is likely that these AI agents are crafted using the works of those they emulate, but the legality of utilizing such content remains questionable and has sparked numerous copyright disputes.
“Our Expert Review agent analyzes the writing a user is crafting, whether it’s a marketing piece or an academic project on biodiversity, and uses our underlying large language model to present expert content that can assist the author in refining their work,” explained Jen Dakin, senior communication manager at Superhuman. “The recommended experts vary based on the subject matter of the text being reviewed. The Expert Review agent does not claim endorsement or active involvement from those experts; it offers suggestions inspired by their works and directs users toward prominent voices they can explore further.”
While someone like King might view the rise of AI as inevitable, concerns arise around the numerous luminaries who wish to protect their intellectual property from being distilled into algorithms. Vanessa Heggie, an associate professor specializing in the history of science and medicine at the University of Birmingham, recently voiced her discontent on LinkedIn, accusing Superhuman of “creating little LLMs” derived from the “scraped work” of both living and deceased individuals, profiting off their “names and reputations.” She shared a screenshot highlighting the availability of analysis from an AI agent modeled after historian David Abulafia, who recently passed away. “Obscene,” Heggie remarked.