Facilitator guide: Deepfakes & Misinformation

Timing for a 60-90 minute class

Segment	60 min	90 min	What happens
Pre-check	5	5	Learners answer the three questions. Keep their answers to compare at the end.
Step 1: Hook	5	8	Read the Ethan story, then take a vote: real or fake, and how would you check?
Step 2: Core idea	10	15	Explain how fakes are made, then run the clue sort in small groups.
Step 3: Why fakes spread	8	12	Teach disinformation and misinformation, discuss the two reasons, then run the sort.
Step 4: Check before you share	8	15	Teach the three moves, then try them on Ethan’s clip. Add the optional demo at 90 minutes.
Step 5: Scenario	8	12	Learners vote on A to D, then discuss the follow-up.
Step 6: Quick check	4	6	Answer the four questions, then go over each explanation.
Step 7: Reflect	7	10	Private writing first, then the discussion prompts.
Step 8: Recap and post-check	5	7	Read the recap, repeat the three pre-check questions, and compare.

Before you start (sensitive spots, what to avoid)

• Use only the invented clips in this module. Don’t show or link a real deepfake, and don’t make one with learners.
• Ethan, Marcus, and Jasper are made up. Don’t name real classmates, teachers, or public figures in examples. If a learner has one of these names, swap it.
• Some learners may have been the target of a fake, or may have shared one. Reflect is private, so don’t ask anyone to read theirs aloud or say what they shared. Point any learner who wants to talk to a trusted adult.
• Keep the tone calm. The goal is habits, not fear.
• Step 5 mentions cheating on a test. Keep the focus on the habit of checking, not on who cheats.
• Many AI tools require users to be 13 or older, and these learners are 11 to 13. Don’t tell anyone to try a tool. Any demo is led by you.
• The clue labels are for today. As of 2026, visual clues are weak, but tools change, so tell learners the habits matter most.
• Name AI products only as examples. Don’t recommend a detector or any other tool.

Discussion prompts

Use the three in Step 7. For extra time, try: “Should apps label AI-made clips? What are the pros and cons?” Show both sides and let learners decide.

Answer notes

• Step 1: There is no single answer. Steer toward who posted it first and whether a trusted source confirms it.
• Step 2 sort: Cards 1 to 3 are strong clues: one new account, no second trusted source, and an original that says something different. The hands, the “looks completely real” card, and the detector tool are weak. Ask learners to defend their picks, and remind them these labels are for today.
• Step 3 sort: Marcus’s clip is disinformation. The friend forwarding it is misinformation. The page posting fakes for clicks is disinformation. Sharing a fake you believed is misinformation. Ask: can one clip be both? Yes. It can start as disinformation and become misinformation when others share it believing it’s real.
• Step 5: D is the strongest choice. A spreads an unconfirmed clip. B and C each help partly. Ask what learners would check first.
• Step 6 answers: B, D, A, C.
• Pre/post check answers: C, A, D.

Further reading

• The Creation and Detection of Deepfakes: A Survey (Mirsky and Lee, ACM Computing Surveys, 2021)