Your next email draft. You’re tweaking a pitch, hit that shiny ‘Expert Review’ button in Grammarly, and boom — Stephen King tells you to punch up the prose. Sounds great, right? Except it’s not him. Not even close. This stunt from Grammarly — now awkwardly rebranded under Superhuman — just nuked whatever trust folks had left in AI writing helpers.
Real people like you and me rely on these sidebar buddies to sound sharper, avoid typos, land jobs. But when they start hallucinating ‘experts’ with your colleagues’ or dead scientists’ names? Kiss goodbye to that blind faith. We’re back to squinting at every suggestion, wondering who’s really whispering in our ear.
Grammarly’s Desperate Rebrand Play
Look, I’ve covered these Valley pivots for two decades. Grammarly, that trusty grammar nag, sniffed the AI gold rush and snapped up Superhuman Mail last June. October rolls around, they’re ditching the old name for ‘Superhuman’ — bold move, or panic? CPO Noam Lovinsky swears the Grammarly brand sticks around, just souped up with AI agents.
But here’s the cynical bit: who’s banking? Not users footing premium subs, that’s for sure. It’s execs chasing that enterprise AI dream, where companies pay fat stacks for ‘agents’ that do… what, exactly?
Expert Review snuck in back in August, pre-rebrand fireworks. A help page — now scrubbed — promised ‘insights from leading professionals.’ Select it, and AI spits suggestions ‘inspired by’ big names, checkmark and all. Carl Sagan on physics prose? Neil deGrasse Tyson schooling your tweet? Cute, until Wired outs them using dead profs in March.
“After careful consideration, we have decided to disable Expert Review as we reimagine the feature to make it more useful for users, while giving experts real control over how they want to be represented — or not represented at all.”
That’s Superhuman’s Ailian Gan, scrambling after the heat.
Why Did Experts Freak Out?
We at The Verge — okay, fine, the original reporters — fed it drafts. Bam: my buddies Nilay Patel, David Pierce, Tom Warren, Sean Hollister pop up. No permission. Advice? Generic drivel like ‘add urgency and intrigue’ under Nilay’s name. Broken source links to boot.
VP Alex Gay dodges: experts show because their work’s public. Duh. But scraping public words to fake endorsement? That’s the sloppelganger — stolen doppelganger — vibe that tanked it.
March 4th, Wired lights the fuse. By March 10th, opt-out inbox launches. Next day? Feature killed. CEO Shishir Mehrotra on LinkedIn:
“we received valid critical feedback from experts who are concerned that the agent misrepresented their voices.”
He calls it a ‘bad feature,’ buried with ‘very little usage.’ On Decoder, Nilay grills him — Mehrotra repeats it’s trash. Apology issued, but LinkedIn roasts him anyway.
Short para for punch: Epic backpedal.
This isn’t new. Remember 2010s content farms? Sites scraped bios, spun SEO slop under celeb bylines. Google nuked ‘em. Or deepfake porn scandals — same vibe, consent zero. Grammarly’s twist? They’re not shady startups; they’re in your browser, daily. Unique insight: this foreshadows AI persona licensing wars. Expect VCs funding ‘expert avatar’ startups, real writers cashing checks to train models. Or lawsuits, à la NYT vs. OpenAI.
Will AI Writing Tools Ever Get Consent Right?
So, what’s next for Superhuman? Mehrotra vows rethink — ‘real control.’ Opt-ins? Paid partnerships? Doubt it; margins matter more than manners.
Users? We’re warier now. That Grammarly sidebar — once harmless — feels like a spy. Enterprise suits might shrug, but freelancers, students? They’re dumping it for rivals like ProWritingAid or plain old Word.
And the money question: Superhuman’s burning cash on agents nobody asked for. Grammarly hit $1B ARR pre-pivot; this PR dumpster fire? Could shave millions. Valley loves ‘move fast,’ but screw the little guy — regulators circling.
Dig deeper. Feature flew under radar months. Why? No fanfare — they knew it smelled off. Subtle disclaimer buried: ‘no affiliation.’ Too late.
Cynical vet take: AI hype’s a house of cards. Companies bolt ‘AI’ labels, pray for moat. But ethics trips ‘em every time. Prediction — bold one — by 2027, we’ll see ‘Verified Expert’ badges as premium upsell. Paywall for trust.
Broader ripple. Tools like Jasper, Copy.ai watching. One wrong scrape, and poof — class actions. Users revolt on Reddit, App Store bombs.
Here’s the sprawler: think about the ecosystem, how Grammarly hooked schools, corps with free tiers, then premium-locked the good stuff; now this ‘expert’ gimmick aimed at pros needing polish, but it backfired spectacularly because nobody wants ghostwritten-by-ghosts advice, especially when the ghosts are alive and pissed, leading to a swift kill-switch that screams ‘we knew better but shipped anyway,’ classic Silicon Valley hubris that I’ve seen crash unicorns from Pets.com to WeWork-wannabes.
But. Hope flickers. If Superhuman nails opt-in with royalties, they pioneer fair play. Unlikely, but hey.
The Human Touch in AI Tools
Strip the spin: we crave smarts, not simulacra. Real editors, not robo-King.
Mehrotra’s Decoder mea culpa? Admits low usage — so why build it? KPI chase, probably.
For real people: double-check that AI nudge. Trust, but verify.
FAQ time.
🧬 Related Insights
- Read more: OpenAI’s Bug Bounty Arms Hackers Against Its Own AI Nightmares
- Read more: Orbital Datacenters: AI’s Escape from Earth’s Energy Shackles
Frequently Asked Questions
What happened with Grammarly’s Expert Review feature?
It generated writing tips under real experts’ names — living, dead, unasked — without permission. Backlash killed it in days.
Is Grammarly now Superhuman?
Rebrand underway, Grammarly lives as a hub inside. Expert Review’s toast, for now.
Should I still use Grammarly?
Basic grammar? Sure. Fancy AI? Proceed with side-eye — consent issues linger.
