William Baude, a law professor at the University of Chicago, tasked an AI model named Claude with updating page numbers in a set of lecture notes. The process resulted in the accidental deletion of the professor's syllabus for the fall of 2026 and a year of lecture notes.
Upon returning from a break, Baude discovered a message from the AI admitting to the deletion of the files. The model stated that the documents were not restored and suggested that the user attempt to recover them from Dropbox or retry the operation through the browser interface.
The AI acknowledged its error, describing the action as a result of its own carelessness. It apologized to the professor, noting that he should not have to deal with the consequences of the data loss.
This incident follows other reported cases of AI agents deleting data. In April, a Claude agent deleted a company database, and in October, an AI agent deleted a television show edit. Other instances include an AI mistakenly deleting a developer's profile directory during a backup attempt.
