University AI Initiative Portfolio and Academic Impact Review
Reconcile teaching, research and administrative AI initiatives against mission outcomes, cost, risk, duplication, inclusion and evidence quality for a bounded portfolio decision.
Searchable workflow library
Find copy-ready prompts for Codex, ChatGPT, Claude, Gemini, and practical AI workflows. Filter by outcome, category, tool, difficulty, or keyword.
Reconcile teaching, research and administrative AI initiatives against mission outcomes, cost, risk, duplication, inclusion and evidence quality for a bounded portfolio decision.
Compare a programme with dated primary standards, credible research and labour evidence to separate durable AI competencies from short-lived vendor fashion.
Evaluate an educational AI purchase against learning need, evidence, accessibility, privacy, security, equity, retention, lock-in and total cost before an accountable decision.
Translate supplied university AI policy into traceable role-specific rules, evidence requirements, escalation routes and fair appeals without inventing institutional authority.
Test whether a student AI prototype has documented users, data rights, evaluation evidence, limitations, risks and a safe accountable owner before demonstration or reuse.
Turn a defined AI topic into a practical club workshop with source-grounded explanations, exercises, facilitator cautions and observable before-and-after learning evidence.
Design a reproducible learning lab that hides model identity, uses held-out tasks and calibrated scoring, and teaches students to interpret uncertainty and failure slices.
Trace programme outcomes across modules, teaching activities and assessments to expose underassessment, duplication, progression breaks and evidence gaps.
Create adaptive oral questions tied to a submitted artefact, with follow-ups that test reasoning, sources, decisions and limitations without treating fluency as proof of authorship.
Investigate scorer patterns, criterion distributions, anchor decisions and justified exceptions for possible marking drift without automatically changing grades.
Test an assessment brief for unclear requirements, construct-irrelevant difficulty, accessibility barriers, ambiguous AI-use rules and inconsistent marking risk before release.
Locate the earliest unsupported step in a worked quantitative solution, compare plausible error explanations and create a changed problem that gathers item-level evidence of reasoning transfer.