March 2026 · · iIA Specialized Courses
Foundations of Critical AI for Libraries (strategy + practice)
From enthusiasm to defensible decisions: criteria, limits and guided practice to adopt AI without losing trust.
Your team establishes a shared language about AI, identifies priority risks (biases, privacy, hallucinations), defines operational controls, and builds a first set of prompts and best practices tailored to your library.
Problems it solves
- Impulsive adoptions without criteria or limits.
- Unsafe use of data and reputational risk.
- Inconsistent results due to lack of guidelines.
- Pilots without metrics or owners.
Results achieved
- Shared criteria to evaluate tools and use cases.
- Usage limits and practical controls (human review, data minimization, traceability).
- 3–5 viable use cases prioritized for piloting.
- Institutional Promptbook v1 and minimum policy (operational draft).
Course structure
What is (and what is not) the AI your library can use today. Types of models, real capabilities vs. inflated expectations. Documented cases in academic and public libraries. Shared vocabulary so the whole team speaks the same language.
Critical AI framework applied to libraries: why AI does not replace professional judgment. Human-centered design principles: human verification, traceability and explainability. Exercise: identify 3–5 real opportunities in your library.
Biases, hallucinations, privacy and reputational risk: what can go wrong and how to prevent it. Workshop: risk and controls checklist. Deliverable: minimum AI usage policy draft (1–2 pages).
Criteria matrix to evaluate AI tools without depending on trends or vendors. Workshop: Institutional Promptbook v1. Executive close with an operational roadmap to start the following Monday.
Frequently asked questions
No. It is strategic and practical for library teams; it does not require programming or prior technical knowledge.
No. The approach is tool-agnostic. You will learn criteria to evaluate any AI tool.
Four documents: Opportunity map, Risk checklist, Minimum AI usage policy, and Institutional Promptbook v1.
Ideally: a representative from management or coordination and one from IT or digital services. Designed for both profiles to work together.
At least 4 hours of individual work to complete exercises and tailor deliverables to your institutional context.
