A little uncertainty.
A better business decision.
A ready-to-run 5–10 minute opener. Designed for student phones or small teams sharing one device. The default pace is about eight minutes.
What students should learn
By the end, students should be able to distinguish a claim about how content was made from a judgment about whether it is fit for business use, recognize limits in their own confidence, and name a verification step before approving content.
Before students arrive
- Upload the website following SETUP.txt. Test it over HTTPS on your actual subdomain.
- Open the site on the projector, choose Host a game, sign in with your private host password, and create a room. Keep this host tab open; do not share the host browser session.
- On a second device, use the join link or room code, submit one answer, and confirm the host sees it. Use a fresh room for class.
- Set 25 seconds per round for the eight-minute version. Students may use nicknames or one nickname per team. No accounts are required.
The eight-minute run of show
| Time | What you do |
|---|---|
| 0:00–1:00 | Students join. Read the opening script below. Explain the four choices and confidence rating. |
| 1:00–6:30 | Run six rounds, about 55 seconds each: 5 seconds to view, 25 seconds to vote, and 25 seconds to reveal and discuss. Ask one student for a reason before each reveal. |
| 6:30–8:00 | Show results. Ask the closing question, hear two responses, and connect the lesson to business responsibility. |
Opening script
“You will see six examples that a business might use. For each one, decide whether the content was created by a human, generated by AI, or combines both. Choose ‘Unsure’ if you do not have enough evidence, and rate your confidence.
For this activity, ‘AI’ means generative AI used to create content. Ordinary scanning, digitization and basic photo processing do not count as generative AI. Judge the whole item shown.
This is an icebreaker, not a test of your intelligence or a reliable AI detection method. After each reveal, we will ask the more useful question: would you approve this for business use, and what would you verify first?”
Six rounds, with answer notes
Spoilers below. Students can access this guide; the game is designed for learning, not secure examination.
1 · Product catalogue image → AI
Origin: A fictional ceramic cup generated with OpenAI image generation for this workshop.
Ask: “How would we confirm the delivered product matches the picture?”
Listen for: compare with a real sample, confirm dimensions and materials, review the image with the supplier, avoid misleading customers.
2 · Earthrise photograph → Human
Origin: William Anders, Apollo 8, 24 December 1968. NASA image AS08-14-2383.
Ask: “What checks remain, even when the photo is authentic?”
Listen for: source, context, credit, permitted use, and avoiding implied NASA endorsement.
3 · Customer service email → AI
Origin: A new AI-written fictional customer service response.
Ask: “Which claims must be checked before pressing Send?”
Listen for: whether the refund happened, whether policy permits it, and whether the 3–5 day promise is accurate. Discuss customer data only in approved systems.
4 · Business insight quotation → Human
Origin: Adam Smith, The Wealth of Nations (1776), Book I, Chapter II.
Ask: “If a chatbot repeated this, who authored the original words?”
Listen for: differentiate original authorship from the tool used to retrieve it; verify wording and attribution in a source.
5 · Earthrise campaign concept → A mix
Origin: A human-captured historical photograph paired with an AI-written caption, assembled into a fictional workshop mockup.
Ask: “Would this make an environmental claim about our business?”
Listen for: evidence for sustainability claims, approval of copy, image attribution, and avoiding an implied partnership or endorsement.
6 · Chatbot replacement proposal → AI
Origin: A deliberately flawed AI-written proposal. The statistics are invented and unsupported.
Ask: “Would a human author make this recommendation any more reliable?”
Listen for: business baseline, cost assumptions, a pilot, error evaluation, customer impact, escalation and human oversight.
Closing script
“Which example would you approve for business use—and what would you verify first?”
“The source matters, but it does not establish quality. Human work can be wrong, and AI work can be useful. A manager’s job is to check the evidence, consider the impact, and remain accountable for the decision.”
Adapt the timing
Five minutes: let students join while settling in; use 20-second voting, give a brief reveal, and save discussion for one final question. Ten minutes: use 30-second voting and invite a reason before each reveal. If internet access is unavailable, project the samples from this guide’s source links or the assets folder, and use a show of hands for the four choices.
Scoring and confidence
A correct origin earns 100 points. All six rounds total 600 points. There is no speed bonus or confidence bonus. “Unsure” earns zero points, but is a legitimate statement of uncertainty. Equal scores share a rank; alphabetical display order does not break a tie. Avoid treating the leaderboard as an academic grade.
Confidence is a self-reflection measure. Each reveal shows how many votes were both highly confident and incorrect, without naming students. Ask what evidence could have changed their judgment.
Sources and teaching rationale
The activity’s emphasis on critical judgment aligns with UNESCO’s AI competency framework for students. For limitations of synthetic-content detection and the role of provenance, see NIST’s digital content transparency report. Sample-specific origins are documented in the provenance record.