Why are writers suing Grammarly's feature?
What happened and why it matters
A recently launched Grammarly capability that presented editing suggestions "inspired by" named authors and experts provoked a sharp backlash and has triggered legal action. The feature surfaced AI-generated feedback framed as coming from real writers and academics without their consent. After reports that the tool attributed its recommendations to identifiable figures, platforms that surfaced the feature — including Superhuman, which had integrated Grammarly’s tool — moved to disable it while companies and affected writers weighed legal options.
The controversy escalated into litigation. A proposed class-action complaint was filed by journalist Julia Angwin, alleging misappropriation of identity and commercial exploitation of writers’ profiles. Other writers and organizations called out the practice as a breach of consent and professional reputation. The suit and public outcry center on two legal and ethical questions: whether a company can commercialize an AI capability that mimics or claims to be modeled on specific living authors without permission, and whether the companies’ disclosures and opt-out options are adequate to protect those authors’ rights.
Key implications
- Reputation risk: Authors and academics objected to their names being used as a veneer for AI outputs they did not produce, creating immediate PR damage for vendors.
- Legal exposure: Misappropriation and likeness claims are now in play, and class-action filings could force wider settlements or changes to product design.
- Industry practice: The episode is accelerating industry conversations about attribution, consent, and the limits on presenting AI outputs as coming from living people.
What to watch next
Companies that build or integrate generative systems will likely tighten policies around using the identities of real people. Regulators and courts will also have to confront how existing privacy and publicity laws apply to generative AI. It’s still unclear how broadly courts will interpret misappropriation in this context, but the immediate result is a pullback of the feature and a renewed debate over consent and attribution in AI.