Educators are rethinking assessment design as students increasingly use AI tools to complete assignments; classrooms are adopting honor‑code disclosures, in‑person checks and project‑based evaluations to preserve academic integrity. Districts and campuses are also weighing data privacy and environmental costs tied to large language models, while teachers experiment with ethical‑use frameworks and AI‑augmented pedagogy. At the same time, AI insiders predict a shift from massive general‑purpose models to smaller, domain‑specific agents that run on local machines and cost far less, according to industry interviews and analysis. Campus IT, research computing and teaching units must prepare for both assessment redesign and a changing technology stack — from cloud‑scale LLM dependence to specialized, on‑prem models that alter procurement, privacy and grading workflows.