Professors at Arizona State University raised concerns that an internal AI platform project using “Atomic” may be scraping course materials without faculty input. The reported friction centers on decontextualized lecture snippets and error-prone summary outputs showing up in the service. Faculty stakeholders say the process may affect academic integrity, intellectual property, and teaching autonomy—especially if course materials are treated as raw data for commercial-style AI tooling rather than governed learning assets. The conflict illustrates how generative AI procurement and experimentation on campus is moving faster than standard faculty review and consent processes. As universities expand AI-enabled teaching tools, this case is likely to inform governance expectations around data provenance, acceptable use, and how faculty expertise is embedded in model deployment and evaluation.