Before the semester ended, the professor used an AI-detection tool on final exams, flagged thirty papers as machine-generated, and gave failing grades. According to coworkers, he acted out of genuine frustration. Over the course of two years, he had observed a change in the writing quality: the same sentence rhythms, neat three-part structures, and odd comfort with em dashes. The feeling is familiar to anyone who has graded student work since 2023. There’s a feeling that something has changed in the classroom, and no one knows exactly how to react.

However, his trusted tools were never as dependable as the advertising claimed. A flagged essay only “might be” AI-generated, according to Turnitin’s own fine print, which makes it odd to base a disciplinary procedure on. Non-native English speakers are more likely than anyone else to be falsely flagged, according to independent research.

The same alerts are occasionally set off by students who use Grammarly for grammar checks, which is something that universities themselves frequently advise. It is a loop reminiscent of Kafka. You are expected to demonstrate your humanity when one piece of software accuses you of being led by another piece of software.

A veteran returning to school on the GI Bill and a first-generation student who had made the dean’s list three times in a row are reportedly among the thirty students in this case. Their attorneys contend that the professor did not thoroughly review the papers that were flagged before assigning grades. The university has so far refused to make a compromise because it is torn between defending its faculty and maintaining its reputation.

It’s difficult to ignore a pattern as you watch this develop. Approximately 90% of the nearly 6,000 academic misconduct cases that Australian Catholic University reported in a single year were related to artificial intelligence. Afterwards, a quarter were dismissed. Some students had their transcripts frozen for months while job offers vanished, such as Madeleine, a young nurse who publicly discussed her situation. The harm is not hypothetical. Rescinded offers, postponed graduations, and the strange humiliation of being accused by a machine are some of the ways it manifests.

Beneath all of this lies a deeper question. Essays are supposedly assigned by universities to teach students how to think. Something crucial is lost if a surveillance system is built around each paragraph a student writes in response to generative AI. Oral exams, in-class writing, and recorded presentations are taking the place of the traditional essay, which some professors are quietly doing away with. It seems to be moving more slowly. It has a more human vibe. It may also be the only option that avoids going to court.

Over the course of the following year or two, the lawsuit will proceed through the legal system. It is possible to reach agreements. Rewritten policies are possible. However, the greater discomfort persists. Teachers are being asked to use tools that don’t fully function and police technology they don’t fully understand against students who are frequently just trying to keep up. Thirty people failed a class they most likely should have passed during that time. At least that part is uncontested.

The post A University Professor Used AI to Detect AI-Written Exams and Wrongly Failed 30 Students. A Lawsuit Followed appeared first on Creative Learning Guild.