Distinguish fluency from reliability
Explain why confident, detailed or persuasive AI-generated language does not by itself demonstrate accuracy, evidence or understanding.
Erasmus+ teacher training course
AI on Trial: Question, Verify, Decide
How can educators turn confident AI outputs into opportunities for investigation and critical thinking?
Help learners question, compare and verify generated content instead of accepting it as an answer.
Target group: Teachers, adult educators, trainers, curriculum coordinators, school leaders and other education professionals from all educational levels.
- Activity type
- Physical mobility
- Duration
- 1 week
- Language
- English
- Equipment
- Laptop required
Critical Thinking and AI Verification for Educators
AI-generated information can sound convincing even when it contains weak arguments, invented sources or missing context. Instead of treating AI only as a tool for producing answers, this course uses its output as material for investigation. Participants learn how AI can create opportunities for learners to question, compare and make informed decisions.
Practical activities focus on separating claims from evidence, checking sources, comparing different accounts and looking beyond generated citations. Participants examine bias, uncertainty, privacy and the limits of automated answers. They also explore how to build a classroom culture where learners explain their reasoning, recognise when evidence is weak and feel able to change their opinion.
Each participant designs and tests a learning sequence in which learners question AI-generated material, check evidence, record their reasoning and use human judgement to reach a supported conclusion.
Learning outcomes
What You Will Learn
Interrogate generated claims
Break an AI response into claims, assumptions, omissions and uncertainties, then formulate questions that require further investigation.
Verify information and trace sources
Use source tracing, lateral reading and claim-evidence comparison to determine whether external evidence genuinely supports a generated statement.
Design AI activities that preserve learner thinking
Create learning sequences that require independent thought, purposeful AI interaction, documented reasoning and revision rather than passive acceptance or copying.
Establish responsible human oversight
Define appropriate boundaries for privacy, accuracy, bias, transparency, intellectual property and professional accountability when using AI in education.
Six-day programme
Course Schedule
Day 1AI answers are claims, not evidence
First Answer, Second Look
Participants examine an instructor-created AI response, record their immediate reaction and then identify which parts are claims, interpretations, recommendations or unsupported assertions.
Confidence Is Not Correctness
Participants compare outputs that differ in tone, detail and certainty, then discuss how polished language, references and technical vocabulary can influence perceptions of reliability.
Map the Claim Chain
Small groups divide a generated answer into individual claims and mark what appears verifiable, uncertain, value-based, incomplete or dependent on missing context.
Day 2How questions and contexts shape outputs
One Question, Several Answers
Participants compare responses produced from variations of the same question and identify how wording, context, role instructions and requested format influence what appears in the answer.
What Is Missing?
Participants examine whose perspectives, experiences, evidence or possible consequences are absent from an output and rewrite the inquiry to expose those gaps.
Document the Prompt Trail
Participants create a transparent record showing the original purpose, information supplied, prompt changes, generated outputs and decisions made by the human user.
Day 3Following claims back to evidence
Citation Chase
Participants investigate a set of instructor-created outputs containing reliable, weak, irrelevant and invented citations, then determine whether each source exists and supports the associated claim.
Lateral Reading Sprint
Participants leave the original page or generated answer to investigate authorship, publication context, independent coverage and evidence from other sources.
Build an Evidence Ledger
Teams create a concise record connecting each important claim with its supporting source, evidence strength, unresolved questions and degree of confidence.
Day 4Bias, framing and persuasive fluency
Polished but Weak
Participants examine an apparently convincing AI-generated argument and identify vague authority, false balance, unsupported causation, selective evidence and rhetorical shortcuts.
Construct the Strongest Counterclaim
Participants ask what evidence could challenge a generated conclusion, build the strongest reasonable alternative and decide whether the original position should be retained, qualified or rejected.
The Boundary Case
Participants discuss situations in which AI assistance may be appropriate, questionable or unsuitable and justify where human expertise, care or accountability must take priority.
Day 5Designing tasks that protect thinking
Think Before the Tool
Participants redesign an activity so learners form an initial interpretation, prediction or solution before consulting an AI system.
Productive Friction
Participants introduce checkpoints that require comparison, explanation, disagreement, source verification or revision instead of allowing a generated answer to end the learning process.
Assess the Reasoning Trail
Participants identify evidence of learning beyond the final product, including question development, source choices, revisions, oral explanation, uncertainty and justified disagreement.
Day 6Course Closure & Cultural Activities
- Reflecting on learning outcomes and key takeaways
- Awarding of Certificates of Attendance
- Cultural excursion and local heritage experience
- Informal networking and exchange of best practices
Course foundations
Recommended Reading and Theoretical Influences
- UNESCO (2023). Guidance for Generative AI in Education and Research.
- UNESCO (2024). AI Competency Framework for Teachers.
- European Commission (2022). Ethical Guidelines on the Use of Artificial Intelligence and Data in Teaching and Learning for Educators.
- Vuorikari, R., Kluzer, S. and Punie, Y. (2022). DigComp 2.2: The Digital Competence Framework for Citizens.
- Wineburg, S. and McGrew, S. (2019). Lateral Reading and the Nature of Expertise: Reading Less and Learning More When Evaluating Digital Information.
- Facione, P. A. (1990). Critical Thinking: A Statement of Expert Consensus for Purposes of Educational Assessment and Instruction.
Completing the course
Evaluation, Recognition and Practical Information
Learning is supported throughout the week through participation, reflection, practical activities and constructive feedback.
Evaluation of Learning Outcomes
- Format
- Continuous assessment based on class participation, practical work and self-assessment.
- Criteria
- Active participation in discussions, reflection, collaborative activities and practical course tasks.
- Procedures
- Trainer observation, participant reflection, feedback and attendance.
Recognition of Learning Outcomes
- Conditions
- Attendance at a minimum of 80% of the course and active participation in the learning activities.
- Recognition
- Learning outcomes are recognised through participant reflection, trainer feedback and completion of the course activities.
- Documentation
- A Certificate of Attendance documenting the course title, dates, venue and learning outcomes.
Learning approach
Practical, Collaborative and Learner-Centred
The course combines hands-on workshops, real-world examples, simulations, guided reflection and collaborative activities. Participants exchange good practices, work in international groups and develop ideas that can be adapted to their own professional context.
Optional cultural, social and networking activities support local engagement, intercultural learning and professional collaboration.
Before you travel
Practical Details
Duration
One week, normally comprising 25 academic hours.
Weekly Schedule
Classes take place from Monday to Friday, in the morning or afternoon. Saturday is reserved for cultural activities.
Final Timetable
The detailed timetable will be sent at least two weeks before the beginning of the course.
Preparation
No special preparation is required unless stated in the course programme. Any required materials or equipment will be communicated before the course.
Certification
Participants who meet the attendance requirements receive a Certificate of Attendance.
Course Fee
The standard fee is €480. See what the course fee includes.
Alternative arrangements: Other course durations and schedules may be arranged on request.
Erasmus+ funding: Course fees and mobility costs may be supported through an eligible sending organisation's Erasmus+ grant. Eligibility and final funding decisions remain with the beneficiary organisation and its National Agency. Read our Erasmus+ KA1 funding guide.
Administrative support: Understanding Academy provides course programmes, learning outcomes, registration documentation and certificates. Participants and sending organisations remain responsible for transport, accommodation and grant management.
Plan your Erasmus+ mobility
Ready to Join This Course?
Check the upcoming confirmed dates or register your interest. We will contact you with information about availability and the next steps.
Need funding information? Read our Erasmus+ KA1 funding guide.