Grammarly Expert Review Scandal: Privacy Risks in AI

The short version

Grammarly, the popular writing app owned by Superhuman, has a new AI feature called "Expert Review" that uses real journalists' and authors' names to make its suggestions seem more credible—without asking permission first. The company isn't apologizing or stopping it; instead, they're letting affected people opt out by emailing them, which critics say puts the burden on victims who might not even know it's happening. This matters to you because if you use Grammarly (over 30 million people do), you're relying on AI that might fake human expertise, eroding trust in the tools you use for emails, reports, and schoolwork.

What happened

Imagine you're writing an email in Grammarly, and it pops up a suggestion like, "Fix this sentence—Nilay Patel from The Verge would say it's too wordy." Sounds legit, right? But journalists from The Verge, Wired, and elsewhere discovered their names were being used this way without their okay. The Verge's reporters tested the feature and saw their own boss's name, Nilay Patel, and colleagues like David Pierce, Sean Hollister, and Tom Warren listed as "experts" giving AI-generated advice.

Grammarly calls it "inspired by" public writings from these experts, not direct quotes or endorsements. But it displays their real names next to the tips, making it look like a real person reviewed your work. When called out, Grammarly (via Superhuman) didn't say sorry or pull the feature. Instead, VP Alex Gay issued a statement saying they've heard the complaints and will let "experts" opt out by emailing expertoptout@superhuman.com. They're "working on refinements," but no big changes yet. Critics like Sean Hollister argue this is backwards: Why should busy authors hunt down an opt-out they didn't know existed? And what about everyone else who doesn't use Grammarly or read tech news?

It's like a restaurant using a famous chef's name on the menu to sell burgers the chef never tasted—without asking. The chef has to email to get their name removed, and most people never find out.

Why should you care?

Grammarly isn't just some niche tool—it's in your browser, phone, and email for millions, checking everything from job applications to love letters. This scandal shows how AI companies are borrowing real people's reputations to boost their products, blurring the line between human smarts and machine guesses. If the "experts" are faking it, your writing advice might be less reliable than you think. Worse, it sets a precedent: What if your name or words get used next time you're quoted online? For everyday users, it chips away at trust—why rely on a tool that plays fast and loose with identities?

On a personal level, this hits your privacy. Grammarly already reads your private docs (with your permission when you install it). Now, it's clear they'll stretch rules on others' data too. Even if you opt out of their data use, they might keep some stats like how often suggestions are clicked, per reports. In a world where AI is everywhere, knowing when it's truly "helpful" versus hype matters for your work, school grades, or professional emails.

What changes for you

Not much immediately, but here's the practical scoop:

If you use Grammarly: Keep using it—the feature is still there, and suggestions might reference unnamed "experts" more now to avoid backlash. But double-check big advice; it could be AI dressed up as human wisdom.
If you're a writer or public figure: Email expertoptout@superhuman.com ASAP if you've been cited online. But it's on you to monitor, which isn't fair.
Privacy settings: Grammarly lets users opt out of data training their AI, but this is about others' names, not yours directly. Review your account at grammarly.com—turn off "Improve my suggestions" to limit data sharing.
No app crashes or price hikes: Your daily use stays the same, free tier included. But expect more scrutiny on AI writing tools; competitors like Microsoft Editor might tighten rules.
Broader ripple: This pushes companies toward "opt-in" consent (ask first), which could mean fewer sneaky features but more pop-ups. For regular folks, it means savvier tool-picking—ask "Does this AI respect names and privacy?"

Long-term, it might make AI writing helpers disclose when they're faking expertise, like a label saying "AI-generated, inspired by public sources." Until then, treat Grammarly tips as helpful hints, not gospel from pros.

Frequently Asked Questions

### Does Grammarly still read my private documents?

Yes, Grammarly accesses your text to give suggestions—that's how it works, and you agree to it when signing up. This scandal is about them using other people's names in those suggestions without permission, not directly stealing your data. To minimize sharing, go to settings and disable data uploads for model training.

### Is Grammarly free, and does this affect pricing?

Grammarly has a free version for basics and paid plans ($12–30/month) for advanced features like this Expert Review. Nothing here changes costs—it's the same app. The controversy might lead to better free privacy controls down the line.

### How do I opt out if my name is being used?

Email expertoptout@superhuman.com with your name and details. It's reactive, not proactive—no list of everyone affected. Grammarly says they're refining it, but for now, that's your main protection if you're a public expert.

### Is this different from other AI writing tools like ChatGPT?

Yes—ChatGPT generates advice anonymously or from its training data without naming real people. Grammarly ties suggestions to specific living authors' names for credibility, which feels more invasive. Others like Google Docs or Microsoft might "learn from public data" too, but this called out the name-dropping.

### Will Grammarly apologize or remove the feature?

Not yet—they're sticking with opt-out and tweaks, no apology or full rollback per their statements. Watch for updates; public pressure from The Verge and Wired might force changes. If you're concerned, switch to alternatives like ProWritingAid.

### Can I trust Grammarly's suggestions now?

Mostly yes for grammar and style—they're solid AI. But "expert" tips might be overhyped. Think of it like a study buddy who's good at basics but name-drops profs without their input—useful, but verify important stuff.

The bottom line

Grammarly's Expert Review fiasco is a wake-up call: AI tools are getting clever at borrowing human cred without asking, and companies like Superhuman are responding with half-measures like opt-out emails instead of real consent. For you, the average user firing up Grammarly for emails or essays, it means pausing to question if that "pro tip from Nilay Patel" is legit—your work relies on trustworthy advice. Demand better by checking privacy settings, opting out if named, and supporting tools that respect identities upfront. This isn't killing Grammarly (it's too useful), but it could make all AI writing apps more honest, protecting your trust and privacy in the process. Stay vigilant—your words (and others') deserve it.

(Word count: 912)

Grammarly's "Expert Review" Controversy: What It Means for Your Privacy and Trust in AI Tools