Superhuman has acquired GPTZero, the AI-authenticity service best known for detecting AI-written text. The GPTZero acquisition was announced on June 23, and the financial terms were not disclosed. It is a notable pairing, because a company that helps people generate text just bought one that detects it.
What GPTZero brings
GPTZero started as a tool that flags whether writing was produced by AI. Since then, it has grown into a broader authenticity platform. Its toolkit now includes hallucination detection, plagiarism checks, and a product called AI Vision that measures how much AI content is spreading online.
That last tool makes a striking claim. By GPTZero’s own count, roughly 16% of the internet is now AI-generated. Numbers like that help explain why an authenticity business looks valuable right now.
Why an AI-writing company wants a detector
Superhuman is part of the Grammarly group, whose tools help millions of people write with AI assistance. The plan is to fold GPTZero’s detection into the Superhuman Go assistant. In other words, the same product that can draft text would also flag AI-written content.
The company also wants to keep GPTZero’s core audience. Teachers and students remain a priority, and Superhuman says the deal will widen distribution of these tools. The pitch is balance: pair AI generation with a way to check authenticity.
The catch: detectors aren’t perfect
There is a real tension here, though. AI detectors have a long record of false positives, sometimes flagging human writing as machine-made. That problem hits non-native English writers especially hard, which makes accuracy more than a technical footnote.
The irony is hard to miss, too. Grammarly has faced criticism for AI features that mimic real writers, and now it owns a tool meant to catch exactly that. The move also fits a wider authenticity push. We have seen it in Deezer’s detector that flags how much of a playlist is AI, and in YouTube’s automatic labels on AI videos.
For now, the strategy is clear enough. As AI text floods inboxes and feeds, Superhuman is betting that proving what is human becomes as valuable as generating what is not. Whether the detection is accurate enough to trust is the question that will decide it.
