Facilitating a Teacher Workshop on AI

This two-hour pack helps educators review one teaching task and leave with a testable redesign. It implements the AI competency pathway through scenarios, source checking, and peer review. It is a proposed SAGE workshop, not an accredited course or a demonstrated training intervention.

UNESCO’s teacher framework includes a human-centred mindset, ethics, AI foundations and applications, AI pedagogy, and professional learning, with acquire, deepen, and create progression levels.¹ The mapping below is an interpretation for this workshop, not UNESCO endorsement or certification.

Prepare the session

Invite participants to bring an assignment brief they are authorised to share, without student work or confidential assessment items. Supply a fictional brief from the worked assessment designs for anyone who needs one. Ask them to name the learning students must still do themselves.

Prepare accessible copies of the scenario cards, worksheet, source-check exercise, and answer key below. Provide these in advance in editable text, with clear headings and readable tables. Arrange written as well as spoken participation, a facilitator contact, and an appropriate break. No AI account is required: all core activities use shared material and peer review.

If demonstrating a live tool, use an approved environment with synthetic inputs, check it beforehand, and retain the static exercise as a backup. The facilitator should know where to refer privacy, accessibility, security, and assessment questions; they need not improvise institutional approvals.

Two-hour schedule

The 120 minutes include instructions, transitions, and discussion. For a larger group, use parallel pairs rather than adding a report from every group. Reduce the scope of the redesign if participants need longer to read or deliberate.

Scenario cards and facilitator notes

All scenarios are fictional. Provide them without the notes for the first discussion, then compare judgments. Accept alternatives supported by the facts; do not treat escalation as proof that a use is approved or prohibited.

Use Responsible AI use for governance and AI for entrepreneurial learning for card C. The point is to connect a decision with purpose, evidence, data, and consequence, rather than classify AI as uniformly beneficial or harmful.

Source-check exercise and answer key

Supply this fictional evidence card:

S1: A voluntary survey received 40 responses. Thirty respondents said they
would like an evening seminar. The course enrols 200 students. No random
sampling or comparison with non-respondents was conducted.

Intentionally flawed AI-style output: “According to S1, 75% of the course wants an evening seminar. This proves changing the timetable will improve learning.” This is a prepared teaching artefact, not an observed model response.

Ask participants to label each claim as supported, unsupported, contradicted, or unresolved, and write a justified correction. Give five minutes to work, five to compare, and ten to discuss and review the key.

Answer key: 30 / 40 = 75% of respondents favour the seminar. S1 does not establish the proportion of the entire course, because the responses are self-selected. It contains no measure or causal comparison of learning. A supported correction is: “Thirty of forty survey respondents favour an evening seminar; wider demand and effects on learning remain unknown.” The citation is real within this exercise but does not support the original claims.

Use the claim-check record to capture the judgment. Reward accurate interpretation, including retaining a supported claim, rather than the number of errors found.

Copyable redesign worksheet

Original activity and learner context:
Learning outcome and thinking students must do themselves:
Proposed student actions and observable evidence:
AI role: AI-free / AI-limited / AI-permitted / AI-integrated
Allowed actions, prohibited actions, and proportionate disclosure:
Approved tool and permitted material, if a tool is needed:
Sources and educator verification:
Accessible participation and equivalent account-free route:
Independent or fresh-task learning check:
Expected preparation, review, and student workload:
Comparison with current teaching:
Success measure, unacceptable failure, and stop condition:
Decision owner and review point:
Peer-review revision and reason:

For peer review, ask: Which student action demonstrates the outcome? What could AI accidentally do in the student’s place? Which claim or output needs checking? Can every student participate? Is the review burden realistic? What finding would lead the educator to stop? The recipient should revise one substantive decision or explain why it remains justified.

Exit task and follow-up

Use a fresh scenario: “A tutor produces excellent solutions, but students cannot solve a comparable problem without it.” Ask each participant to propose one change and one learning check, explaining their connection to the outcome. An appropriate response might reduce the help given, require an initial attempt, and check a new independent problem. Satisfaction alone does not answer the learning question. This short response samples judgment; it does not certify competence across the framework.

Within two weeks, participants test their draft with synthetic examples or a colleague, complete the pilot record, and report one failure and resulting revision. Involving students requires the usual approvals and access arrangements. Use the next meeting to inspect evidence rather than only share impressive outputs.

Retain workshop worksheets only for an announced professional-learning purpose in approved storage. Do not collect full chats or student records. Evaluate the workshop through the quality of redesigned tasks and later application as well as participant experience. Develop real case reports only when implementation, permissions, results, and limitations are documented.

References & Footnotes