School Surveillance Software Flagged a Student's Essay About 'To Kill a Mockingbird'
Records from three districts show GuardClass's AI flagged literature homework, LGBTQ support searches, and a chemistry study guide as threats.
This is synthetic placeholder content. All companies, people, and events in this article are fictional.
A tenth grader in a suburban district spent two hours in the principal’s office in March because GuardClass, an AI monitoring tool installed on her school-issued laptop, flagged her essay on To Kill a Mockingbird for “violent ideation.” The flagged passage was a quote from the novel.
Incident records from three school districts, obtained through public records requests, show the episode was not unusual. Of 1,412 alerts generated by GuardClass across the districts in one semester, administrators marked 1,287 — 91 percent — as false positives. The flagged material included a chemistry study guide (bomb-related keywords), searches for an LGBTQ youth hotline (categorized “self-harm risk”), and a college application essay about surviving a house fire.
Alerts as a business model
GuardClass sells districts on a simple promise: its AI reads everything students type, search, and write on school devices, and alerts staff before a tragedy. The company’s marketing claims its software has “intervened in thousands of crisis events.” The records suggest a different mechanism: flood schools with alerts, and let overwhelmed staff sort signal from noise.
A school counselor in one of the districts, who requested anonymity because she was not authorized to speak, said the alert queue had changed her job. “I became a content moderator,” she said. “Every morning, forty flags. Thirty-nine are homework.”
The students adapt
The most consistent finding in the records may be behavioral: students learn they are being watched and route around it. Counselors in two districts reported that students stopped using school devices to search for mental health resources — the exact searches the software claims to protect.
A digital rights researcher who reviewed the records for Dead Pixel called this the “chilling effect on the people the product claims to save.”
GuardClass declined an interview. In a statement, the company said its technology “errs on the side of student safety” and that final judgments “always rest with trained school staff” — the same staff its alert volume buries.