AI in Software Testing: Teaching Tips

Start with the Why
- Begin by discussing the limitations of traditional testing and how AI fills the gap.
- Emphasize where and why AI fits into a QA workflow, not just how to use it.
Use Live Demos Frequently
- Demonstrate test generation by typing prompts live into ChatGPT or Copilot.
- Show students how prompt changes impact the output.
Encourage Iteration, Not Perfection
- AI-generated output is a starting point, not a final product. Reinforce the need for human review.
- Encourage students to tweak prompts and compare results.
Guide Critical Thinking
- Ask students: “Would you trust this test? What would you do differently?”
- Discuss potential risks of hallucinations or logic errors in AI-generated code.

Allocate Time for Exploration
- Students should test and refine AI prompts—build in reflection time after each lab.
Encourage Pair Programming
- Pairs or small groups promote discussion about AI reliability and prompt strategy.
Use the Discussion Questions
- At the end of each lab, facilitate a short group conversation based on the “Discussion and Sharing” section.
Point Out Ethical Considerations Early
- Begin seeding conversations about authorship, reliability, and limitations from the start.

Issue	How to Address
AI-generated tests don’t run	Help students debug syntax or framework mismatches (e.g., CommonJS vs ESM in Jest).
Students trust AI output too quickly	Require a review step: have them explain why a test is valid and what it misses.
Flaky test behavior	Use this as a teachable moment—jump to the Flaky Tests lab to reinforce key lessons.
Students stuck on prompts	Encourage them to be specific and structured. Show examples from earlier labs.