AI writing detectors spread in schools and publishing despite doubts

In short: Tools that claim to detect AI-written text are being used more widely, even as many experts and institutions warn they are not fully reliable.

What's going on

AI writing detectors are becoming common in schools and some publishing settings. A Center for Democracy and Technology survey found 43 percent of US teachers in grades 6 to 12 used AI detectors regularly in 2024 to 2025. Some universities using Turnitin, a popular plagiarism checker, found its AI detection feature was turned on automatically when it launched in 2023.

Unlike older plagiarism tools that look for matching text online, AI detectors use software to guess if writing “sounds” like it came from an AI system. It is more like a lie detector for writing than a simple copy check. These tools look at patterns like tone, sentence structure, and word choice, but those patterns can also show up in perfectly normal human writing.

The Verge points to real world fallout from accusations. Publisher Minotaur dropped a reported $2 million book deal over concerns about AI use, which the author Jerry Falade denies. Students have also sued after being accused of using AI, including cases involving Yale and Adelphi University.

What to watch

Some schools, including Yale, Johns Hopkins, Vanderbilt, and Georgetown, have limited or disabled AI detection tools, and MIT warns that “AI detectors don’t work.” A 2023 Stanford study found detectors were more likely to falsely flag writing by non native English speakers, and other reporting has raised concerns about neurodivergent writers too. As platforms add “this looks like AI” style features, more people may get labeled based on a guess, not proof.

Source: The Verge AI