AI, Media Literacy and Agentic Workflows for Educators

Overview

Most staff training on AI stops at "here is a chatbot, try prompting it." This course goes further, because the technology has: agentic systems that plan, use tools and carry out multi-step tasks are already reaching schools and administrators, and the judgement educators need is different from writing a good prompt. You work with predictive, generative and agentic systems side by side, product-neutral throughout, grounded in EU AI ethics guidance, the UNESCO AI Competency Framework for Teachers and the NIST AI Risk Management Framework.

The week is built around practical design work, not lectures about AI. You map a repetitive task from your own institution, design a bounded agentic workflow for it with clear permissions and a human approval gate, write a verification checklist, and run a risk and impact review before anything would actually be deployed. You also work the other side of the same problem: recognising fabricated or manipulated content, teaching media literacy for an age of synthetic images and video, and knowing which tasks should never be handed to a machine at all.

You leave with a concrete output — a bounded agentic-workflow design, an AI-use and verification protocol, or a draft organisational AI-use policy — matched to what your school or organisation actually needs next.

Who it works best for

Educators who already use a generative AI assistant reasonably confidently and now need to think about agentic systems, institutional policy, or how to defend an assessment and a verification process to colleagues and parents. Staff drafting their organisation's first AI-use policy will find the whole week directly useful.

Learning outcomes

• Explain the practical difference between predictive, generative and agentic AI systems, without relying on any single vendor’s terms.
• Formulate goals, context and constraints precisely enough that an AI system’s output is predictable and checkable.
• Design a bounded agentic workflow with defined tools, permissions, stopping conditions and human approval gates.
• Build a verification checklist for AI-assisted outputs, and recognise fabricated, incomplete or manipulated content, including synthetic media.
• Apply EU AI ethics guidance, the UNESCO AI Competency Framework for Teachers and the NIST AI Risk Management Framework to a concrete classroom or workflow decision.
• Draft an organisational AI-use policy that states clearly what must never be delegated to an automated system.

Day-by-day programme

1. Day 1 — Foundations: predictive, generative, agentic

   Welcome and needs mapping. Foundations of AI in plain language, and where educators already use it. A working distinction between predictive, generative and agentic systems, tested against real examples. First exercise: formulating a goal, context and constraints precisely enough that you can judge whether an output actually meets them.

2. Day 2 — Generative AI for planning and teaching

   Using generative AI for lesson design, differentiation and resource adaptation, on material you bring yourself. Practical work on verifying outputs, sources and calculations before they reach a classroom. Recognising fabricated or misleading content, and an introduction to media literacy in an age of synthetic images, audio and video.

3. Day 3 — How agents work

   How agentic systems actually work: planning, tool use, retrieval and multi-step task completion, explained without vendor jargon. You map a repetitive task from your own institution and identify which steps could sensibly be automated. Designing a first agentic workflow with defined inputs, tools, bounded permissions and a stopping condition.

4. Day 4 — Oversight, risk and verification

   Adding human-in-the-loop approval points, and building a verification checklist for multi-step outputs. A risk and impact review before anything goes live, and fallback procedures for missing data or unreliable results. Bias, accessibility, privacy, data minimisation and copyright, read against EU AI ethics guidance, the UNESCO framework and the NIST AI RMF.

5. Day 5 — Assessment, policy and transfer

   Assessment design and academic integrity once AI is part of the picture, and what honest disclosure looks like in practice. Deciding, subject by subject, what should never be delegated. A side-by-side comparison of manual, AI-assisted and agentic approaches to the same task, then drafting your institution’s AI-use policy and action plan.

Who it is for

• Primary and secondary teachers of any subject
• VET trainers and adult educators integrating AI into their practice
• School leaders and coordinators drafting institutional AI-use policy
• Library, cultural-centre and community-learning staff, and professionals supporting digital inclusion

Methodology

No single commercial platform is required: demonstrations and exercises are built around transferable concepts, open formats and workflows you can adapt as tools change. Each day combines short input, guided practice and structured design work — mapping a workflow, building one with approval gates, writing a verification checklist, comparing manual, AI-assisted and agentic approaches to the same task, and reviewing risk before anything is deployed. A pre-course questionnaire and a post-course evaluation keep the week aligned with your own institution’s needs.

Erasmus+ priorities addressed

• Digital transformation
• Inclusion and diversity

What's included

• All course sessions and materials, in digital form and reusable under an open licence
• Pre-course questionnaire and preparation pack
• Certificate of attendance stating contact hours
• Europass Mobility support
• Cultural activity and guided visit
• Local contact person for the duration of the mobility