On Wednesday, March 11, 2026, Superhuman (the parent company of Grammarly) officially disabled the controversial “Expert Review” feature following a fierce backlash from the global writing community.
The feature, which launched in late 2025, used AI to mimic the writing styles and “editorial voices” of hundreds of famous authors, journalists, and academics—often without their knowledge, consent, or compensation.
The Controversy: “Digital Impersonation”
The backlash reached a boiling point in early March 2026 when high-profile writers discovered they were being featured as AI “agents” within the Grammarly sidebar.
- Identity Appropriation: Critics like Kara Swisher and Casey Newton condemned the tool as a “monetization of real identities.” Users were receiving feedback from AI clones of living journalists, such as Nilay Patel (The Verge) and Julia Angwin (investigative reporter).
- “Digital Necromancy”: The tool even featured simulations of deceased figures, including Stephen King (impersonated while alive), Carl Sagan, and the historian David Abulafia, who passed away just weeks earlier in January 2026.
- Misleading Design: In platforms like Google Docs, the AI suggestions appeared in blue bubbles nearly identical to real human comments, leading many users to believe they were receiving genuine human critique.
- Technical Hallucinations: Testing revealed the tool would offer “expert feedback” from Stephen King even on nonsensical lorem ipsum text, and many of the citations linked to outdated, broken, or spammy websites.
Legal & Financial Fallout
| Action | Detail |
| Class-Action Lawsuit | On March 11, journalist Julia Angwin filed a class-action suit alleging Grammarly “misappropriated” the identities of hundreds of professionals for profit. |
| Monetization Concerns | The feature was a cornerstone of Grammarly’s $12/month “Pro” subscription, leading to accusations of “authority laundering”—selling the prestige of experts without paying them. |
| Institutional Risk | Academics warned that the tool could encourage students to “play the algorithm” rather than learn original thought, causing some universities to reconsider their enterprise deals with Grammarly. |
The “Mea Culpa”: CEO Response
In a LinkedIn post on March 11, Superhuman CEO Shishir Mehrotra issued a formal apology:
“We received valid critical feedback from experts who are concerned that the agent misrepresented their voices… After careful consideration, we have decided to disable Expert Review while we reimagine the feature.”
The company has pledged to rebuild the tool with a focus on explicit consent and controllable identity, potentially moving toward a licensed model where writers can opt-in and be compensated for their digital likeness.
What Happens Now?
- The “Opt-Out” Email: Before the shutdown, Grammarly offered an opt-out email for experts, which was widely criticized for being reactive rather than proactive.
- The AI Rebrand: This incident is seen as a major hurdle for Grammarly’s recent pivot to being an “AI Agent” company. Analysts suggest the company may now pivot toward generic “personas” (e.g., “The Academic,” “The Copy Editor”) rather than specific real-world individuals.
