AI can generate study notes and practice questions in seconds, but how can students know whether the answers are supported by their course material or whether important syllabus topics have been missed?

In this session, I will explore the design of SyllabusCheck, an open-source tool for auditing AI-generated study materials against a syllabus and source documents. We will examine evidence-linked question checks, syllabus coverage analysis, duplicate detection, and human review of flagged issues.

The session will demonstrate how to evaluate such a tool using a labelled set of valid and intentionally flawed questions, including the limitations of automated verification. Attendees will leave with a practical approach to making AI-assisted learning resources more transparent, testable, and trustworthy.