Grammarly Disables "Expert Review" After Backlash From Writers and Scholars
Grammarly has turned off its Expert Review AI feature after public criticism from academics, journalists, and working writers. The tool had generated feedback in the style of named experts-including some who are deceased-without their consent. The company says it will redesign the feature so experts can control whether and how they're represented.
What happened
Expert Review let users pick a scholar, editor, or journalist and receive AI feedback modeled on that person's voice. Once people realized the system referenced real names and mannerisms, pushback was swift.
"Over the past week, we received valid critical feedback from experts who are concerned that the agent misrepresented their voices," Superhuman CEO Shishir Mehrotra wrote on LinkedIn. He added that scrutiny improves products and said the company is taking it seriously.
Before the change, experts had to opt out to avoid inclusion-an approach many called unacceptable. After criticism, the feature was disabled and is being reworked to give experts clear control.
James Bareham wrote: "I'm no lawyer, but I think 'We're going to keep stealing your stuff until you tell us you don't want us to steal your stuff' isn't quite the defense Grammarly thinks it is-at least not in the court of public opinion... I canceled my pro account today."
Author and editor Benjamin Dreyer mocked the policy, saying he might "take advantage of their bountifully generous opt-out offer," and would be happy to cause "a few moments' worth of agita" for anyone responsible.
Ailian Gan, Grammarly's director of product management for agents, said feedback showed the feature "missed the mark" and promised transparency as the company works to get it right.
Why this matters for working writers
- Your voice is your equity. If platforms can mimic it without consent, your differentiation erodes.
- Opt-out policies shift the burden onto you. Opt-in is the ethical standard when identity and style are involved.
- Using deceased scholars shows how far style simulation can go-and why clear boundaries are needed.
- Editors and clients may assume "good enough" AI feedback replaces paid expertise, pressuring rates and timelines.
- Reputation risk cuts both ways: having your name approximated by a model can create confusion or misattribution.
Practical steps to protect your voice
- Audit your tools. Check AI settings, opt-out options, and data-use policies for any writing or editing platform you use.
- Search yourself. Look for any AI "personas," prompts, or presets using your name or distinctive style without permission.
- Set boundaries in public bios and portfolios. State your policy on AI training and style imitation in clear language.
- Update contracts. Add clauses that prohibit training on your work, style simulation, or using your name for AI features without written consent.
- Watermark your work subtly (phrasing, structure, or metadata) and keep dated drafts. It helps establish provenance in disputes.
- Speak up early. Public posts and formal complaints get attention, and collective pressure moves product roadmaps.
If you rely on Grammarly today
- Confirm the Expert Review feature is disabled in your account and team settings.
- Review Grammarly's data, style, and personalization settings. Limit any features that store or learn from your drafts.
- For client work, get written approval before using AI feedback that could affect tone, confidentiality, or sourcing.
The bigger thread to pull
This isn't just about one feature. It's about consent, attribution, and the business model of style-as-a-service. Writers should expect more tools to test these boundaries-and be ready with policies, contracts, and a public stance.
If you want structured guidance on using AI without giving up your voice, see AI for Writers.
Your membership also unlocks:
