🔗Stalecollected in 16m

Grammarly Sued Over AI Expert Review

Grammarly Sued Over AI Expert Review
PostLinkedIn
🔗Read original on Wired AI

💡Grammarly sued for AI faking expert advice—key ethics/legal warning for AI builders.

⚡ 30-Second TL;DR

What Changed

Class action lawsuit filed against Grammarly over AI feature.

Why It Matters

This case underscores legal risks of AI impersonation, urging companies to secure consents for real identities. It may prompt industry-wide audits of similar features and influence future AI ethics guidelines.

What To Do Next

Audit your AI outputs for unauthorized attribution to real personas to mitigate lawsuit risks.

Who should care:Developers & AI Engineers

🧠 Deep Insight

Web-grounded analysis with 6 cited sources.

🔑 Enhanced Key Takeaways

  • Grammarly's Expert Review feature, launched in August 2025, used AI-generated personas based on scraped publicly available works from real scholars and authors—including deceased academics like historian David Abulafia (died January 2026)—without consent or disclosure of AI generation[2][3].
  • The feature employed persona prompting techniques rather than full fine-tuned LLMs, allowing Grammarly's base AI to adopt expert characteristics and writing styles to generate feedback, raising questions about personality appropriation for commercial use[2].
  • Over 40 employees from Google and OpenAI, including Jeff Dean (Google's chief scientist), filed an amicus brief supporting Anthropic's concurrent lawsuits against the Pentagon over supply chain risk designation, signaling broader industry concern about government-corporate AI policy conflicts[1].
  • Grammarly offered an opt-out mechanism via expertoptout@superhuman.com in response to criticism, but the feature's fundamental design—presenting AI feedback as expert endorsement without actual expert involvement—remained ethically contested[1][3].

🛠️ Technical Deep Dive

  • Persona prompting: A technique that injects expert characteristics and published work context into a base LLM's prompt, enabling specialized feedback without model fine-tuning[2]
  • Data sourcing: Grammarly used web scraping bots to aggregate publicly available works by referenced experts to inform persona prompts[2]
  • Feature architecture: Expert Review appears as a sidebar widget in Grammarly's main writing assistant, surfacing revision suggestions framed as originating from subject matter experts[3]
  • Disclaimer mechanism: Grammarly included a notice stating 'References to experts in Expert Review are for informational purposes only and do not indicate any affiliation with Grammarly or endorsement by those individuals or entities'—though this appeared insufficient to address consent violations[3]

🔮 Future ImplicationsAI analysis grounded in cited sources

Persona prompting without consent may face regulatory classification as identity misappropriation, similar to deepfake legislation.
The use of real individuals' names and reputations to generate commercial feedback without permission creates legal exposure under emerging AI identity protection frameworks.
AI writing assistants may face mandatory expert consent requirements, paralleling music licensing models.
The backlash suggests future products will need explicit permission from referenced individuals or their estates before deploying persona-based features.
Scraping-based AI training will likely trigger stricter copyright enforcement, following Apple's shadow library litigation.
Grammarly's use of publicly available but copyrighted academic works to train personas mirrors practices now under legal challenge across the industry[2].

Timeline

2025-08
Grammarly launches Expert Review feature as part of broader AI-powered writing assistant updates
2026-01
Historian David Abulafia dies; Grammarly continues using his name and work in Expert Review without disclosure
2026-03-05
Public backlash intensifies as academics and journalists discover Expert Review uses their names without consent; media coverage begins
2026-03-06
Grammarly acknowledges criticism and announces opt-out mechanism via expertoptout@superhuman.com
2026-03-09
Casey Newton publishes critical review in Platformer; Grammarly declines interview request from CEO Shishir Mehrotra
2026-03-11
Grammarly shuts down Expert Review feature following class action lawsuit filing
📰

Weekly AI Recap

Read this week's curated digest of top AI events →

👉Related Updates

AI-curated news aggregator. All content rights belong to original publishers.
Original source: Wired AI