Dartmouth’s provost resigns after a review of undisclosed AI use, Minnesota learning scientists warn careless AI use can undermine learning, and Cornell pairs broad AI access with required training. This week’s question: what can be delegated to AI, and who stays accountable?
Cal State’s $39 million OpenAI contract meets student and faculty resistance, an AACSB survey finds AI is deans’ widest gap between priority and preparedness, and a review of 31 studies sets six principles for where AI help ends and student work begins.
Georgetown’s ChatGPT pilot shows what faculty can lose when personal AI accounts move into enterprise workspaces, a Kazakhstan study links AI governance readiness to students’ intentions to stay enrolled, and federal grants fund AI capacity at UTPB, USF, Clark Atlanta, and Jacksonville University.
Oregon’s NVIDIA faculty program stalls with $1.5 million unspent, new NBER research finds no AI unemployment spike for 2026 graduates, and a study links faculty AI technostress to weaker professional identity. This week’s theme: access to AI is easy; building the capacity to use it well is not.
UNESCO ministers, Universities UK, SUNY, and Utrecht University moved AI governance from guidance to formal policy this week, while new research on disclosure and assessment literacy sharpens the case for institutional accountability over access alone.
Two questions run beneath this week’s AI signals: whether institutions can keep assessment standards consistent, and whether their faculty, curriculum, and computing infrastructure are ready for the work AI now demands.
As AI makes finished products weaker proof of learning, institutions are redesigning assessment around evidence of thinking, while AI fluency spreads well beyond computer science.
