Generative AI in university administrations. A consideration of fit, management requirements, compliance controls and capacity

The choice of a deployment model for generative artificial intelligence is becoming a strategic decision, also for university administrations, that goes beyond purely technical functional characteristics. This paper examines three deployment approaches: an externally managed solution as a commercial product, an open-source platform, and a possible in-house development by the institution itself, using a literature-based requirements framework. It is demonstrated that compliance control and compliance capacity must be considered as independent factors in the context of university administrations. Finally, implications for university management and decision-makers are derived using an institution-type-specific decision matrix.