Grammarly's CEO Defends AI Suggestions Framed as Real Writers' Advice
Superhuman CEO Shishir Mehrotra faced direct questioning about a feature that used the names of real journalists, novelists, and academics to frame AI-generated editorial suggestions. The company shut down the "Expert Review" feature after eight months following backlash from writers whose names appeared without permission.
The controversy surfaced when journalists noticed Grammarly was offering writing improvements attributed to specific people. A suggestion might appear as coming from Stephen King or Carl Sagan, with checkmarks suggesting official endorsement. Technology journalist Julia Angwin filed a class-action lawsuit against Grammarly seeking damages.
During a podcast interview with Verge editor-in-chief Nilay Patel, Mehrotra acknowledged the feature failed on both sides. "It wasn't good for experts, it wasn't good for users," he said.
The Original Pitch
Mehrotra explained the feature's inspiration came from how users already interact with AI. People ask ChatGPT and Claude questions like "What would Nilay think about my writing?" by naming a person they admire. The company wanted to formalize that approach.
When developing the strategy, Mehrotra said he spoke with prominent YouTubers and book authors. They described struggling to maintain connections with audiences between releases. "It'd be really amazing to develop an ongoing connection with my fans," one author told him, according to Mehrotra's account.
The execution didn't work. "It's really hard to distill what you would do as an editor based on the outcome of your published work," Mehrotra said. "It's really hard for AI to do that."
The Payment Question
When Patel asked directly-"How much do you think you should pay me to use my name?"-Mehrotra pivoted to attribution and links. He said experts trade on the internet by publishing content and hope others link to and attribute them.
Mehrotra distinguished between attribution and impersonation. "Every mention was very clearly, 'This is inspired not only by this person, but also inspired by a specific work from this specific person, with a clear attributed link,'" he said.
He dismissed the lawsuit claims as "without merit," arguing the feature didn't constitute impersonation.
A Broader Issue
The dispute reflects how prompt engineering works in practice. Telling an LLM to "write like Hemingway" shapes its output by invoking a specific person's style. Whether that creates legal obligations remains unsettled.
Patel raised a larger concern: "People don't understand the difference between copyrights and trademarks and names and likeness. AI is collapsing those differences faster than ever before."
For writers considering AI tools, the episode highlights unresolved questions about how AI for writers should handle attribution and compensation when it references real people's work or style.
Your membership also unlocks:
