Workshop 2: Detecting AI-Generated Misinformation
Title: Outsmarting AI-Driven Political Misinformation
Time: 60-75 minutes
Objective: Students will learn how to detect AI-generated images, videos, and online content using visual, contextual, and metadata cues.
1. Agenda
| Time | Activity | Description |
|---|---|---|
| 0–10 min | Guess the Fake | Show slides 4–5 (Obama & Pope images). Ask: “Which image is real? What makes you suspicious?” Discuss hands, ears, and lighting. |
| 10–20 min | Direct Instruction | Go through slide 3 (types of implausibilities). Teach how to look for visual and cultural inconsistencies. |
| 20–35 min | Image Analysis Station | Students rotate between stations with different AI-generated vs. real image sets (slides 6–8). They document their detection strategy on a worksheet. |
| 35–45 min | Video Verification Challenge | Use slide 9 (video source/date/location) and break into groups. Show a short clip and ask students to identify how to verify its authenticity. |
| 45–55 min | Tool Demo | Introduce browser tools and fact-checking websites: TinEye, InVID, Snopes, Piyaoba (slides 10–11). Let students test one on their phones/laptops. |
| 55–60 min | Reflection | Group discussion: “Which tool or skill was most useful? How might you use this outside of class?” |
* Lesson plan developed as part of a fellowship project with generous support from the UC National Center for Free Speech and Civic Engagement. Licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
** Not sure how to use this lesson plan? Ask a librarian on your campus or contact Ding at yi.ding@csun.edu