AI in Libraries: The Hidden Privacy Crisis

In a national study I conducted for my doctoral dissertation of how university libraries are implementing AI systems, one library leader shared a distressing discovery. Not long after installing a system, the library learned that the AI was building user profiles that stored students’ intellectual behavior far beyond the library’s standard data retention period. “We had no idea it was happening,” the librarian told me. In another interview, a high-level decision-maker at another library shared, “Our privacy policy, last updated in 2018, addressed traditional database searches. It never contemplated how a machine learning system might use seemingly innocuous data to infer patron interests or identities.”