Grammarly, the writing assistant used by over 40 million people daily, is being sued in a class action lawsuit that alleges the company misled customers about its AI-powered Expert Review feature. The complaint, filed in the Northern District of California, claims Grammarly marketed the service as providing access to human experts when much of the work was actually performed by AI. Not a great look for a company built on trust in language.
The lawsuit was first reported by Wired, and it centers on a feature Grammarly introduced as part of its premium offerings. Expert Review was pitched as a way for users to get their writing checked by professional human editors — people with credentials and subject-matter expertise. Users paid extra for it. The problem, according to the plaintiffs, is that Grammarly increasingly relied on AI to handle those reviews without adequately disclosing the substitution.
That’s the core allegation. Bait and switch.
According to the complaint, Grammarly’s marketing materials and in-app descriptions gave users the impression they were paying for human oversight. The service promised feedback from “experts” who could catch nuances that automated tools might miss — tone, clarity, logical flow, discipline-specific conventions. For professionals drafting high-stakes documents, that kind of assurance matters. But plaintiffs say the reality didn’t match the promise, and that AI was doing the heavy lifting behind the scenes while Grammarly collected premium fees.
The timing here is telling. Grammarly has been aggressively integrating generative AI across its product line. In 2023, the company rolled out GrammarlyGO, its generative AI assistant, and has continued expanding AI capabilities into 2024 and 2025. The company has positioned itself as an AI-first communication platform, partnering with enterprise clients and embedding its tools into workflows across industries. So the tension between “AI-powered” and “human expert” was arguably baked into the company’s own strategic direction.
Grammarly has not yet issued a detailed public response to the lawsuit’s specific claims. The company has generally maintained that its products combine AI technology with human expertise where appropriate. But the lawsuit argues that distinction was never made clear to paying customers — and that’s where the legal exposure sits.
Class action complaints like this one often hinge on how a reasonable consumer would interpret marketing language. If Grammarly’s promotional copy said “expert review” and a user reasonably understood that to mean a human being would read their document, then swapping in an AI system without disclosure could constitute deceptive business practices under California consumer protection law. The plaintiffs are pursuing claims under the state’s Unfair Competition Law and Consumer Legal Remedies Act, among others.
This isn’t an isolated phenomenon. The broader AI industry is grappling with transparency questions as companies race to automate services that were previously human-delivered. Amazon has faced scrutiny over its “Just Walk Out” technology, which was reported by Gizmodo to rely heavily on human reviewers in India rather than pure computer vision. The pattern is consistent: companies market AI capabilities while quietly maintaining human backstops, or vice versa. Either way, the gap between what’s advertised and what’s delivered creates legal risk.
For enterprise customers, the implications are significant. Grammarly claims over 70,000 teams use its platform, including companies in finance, healthcare, and government. If those organizations were routing sensitive documents through a feature they believed involved vetted human reviewers — and instead got AI processing — the data handling and confidentiality questions multiply fast.
And then there’s the pricing issue. Expert Review wasn’t free. Users on Grammarly’s premium and business tiers paid for access to this feature specifically because it promised something beyond what the standard AI grammar checker offered. If AI was performing the reviews, plaintiffs argue, customers were essentially paying a premium for a service they were already getting in a different wrapper.
Short version: they paid for a human. They got a bot.
The case is still in its early stages, and class certification hasn’t been granted yet. These lawsuits can take years to resolve, and many settle before reaching trial. But the filing itself sends a signal to the industry. Companies blending AI and human services need to be precise — almost painfully so — about what customers are actually receiving. Vague language about “expert” involvement won’t cut it if the expert is a large language model.
So what should professionals watch for? First, read the fine print on any AI-assisted service that claims human involvement. Second, if you’re an enterprise buyer, ask vendors directly: what percentage of this service is automated versus human-reviewed? Get it in writing. And third, expect more lawsuits like this one. As AI capabilities expand and companies look to cut costs by automating previously human-delivered services, the gap between marketing and reality will keep generating legal action.
Grammarly built its brand on making communication clearer and more honest. The irony of being sued for unclear and potentially dishonest communication about its own product is hard to miss.
WebProNews is an iEntry Publication