AI Talent Scout: Can the Machine Be Trusted?
Log in to saveStudents form an accountability team reviewing an AI system that sorts audition clips for a talent agency. They build a simple rule-based system and a small trained classifier, then test both for accuracy and bias.
Choose your version
Every level teaches the same core idea — pick the one that fits your time, budget, and energy today.
Setup
Print or handwrite the four AI Talent Scout: Can the Machine Be Trusted? challenge steps and place common paper supplies on one table. Prepare one small sample, dataset, scenario, or recycled-material model drawn directly from the project description; no device, paid tool, room decoration, or purchased kit is required.
Activity
- Read the AI Talent Scout: Can the Machine Be Trusted? challenge aloud, then build a non-AI rule system for sorting audition clips, then train a small classifier on safe student-created gestures or sounds.
- Use paper, index cards, recycled objects, or facilitator-provided sample data to test both systems on unseen examples and record accuracy, false positives, and false negatives on a shared dashboard.
- Run one tabletop trial and change the lighting, background, voice, or movement and retest — the model that performed well in training often performs worse here.
- Finish by deliver a go/revise/reject recommendation and design a transparent audition process that keeps a human responsible for the final decision; record the evidence that supports the decision.
Done when: The group completes a defensible AI Talent Scout: Can the Machine Be Trusted? solution, records at least one test result or comparison, changes one claim or design in response, and supports its final decision with that evidence.
Materials
Laptop or tablet, simple gesture/sound recording setup, printed data-tracking sheets
Shown for the Quick Start version — switch tabs above to see what changes.
Disclosure: Materials links below go to Amazon. As an Amazon Associate, Creative Dad Company earns from qualifying purchases at no extra cost to you.
- Shop
Clipboard for data tracking
Steps
- Build a non-AI rule system for sorting audition clips, then train a small classifier on safe student-created gestures or sounds.
- Test both systems on unseen examples and record accuracy, false positives, and false negatives on a shared dashboard.
- Change the lighting, background, voice, or movement and retest — the model that performed well in training often performs worse here.
- Deliver a go/revise/reject recommendation and design a transparent audition process that keeps a human responsible for the final decision.
Standards it satisfies
NGSS · HS-ETS1-3
Evaluate a solution to a complex real-world problem based on prioritized criteria and trade-offs that account for a range of constraints, including cost, safety, reliability, and aesthetics, as well as possible social, cultural, and environmental impacts.