Below we provide some advice and methods on how to incorporate GenAI ethically and responsibly into your courses. More advice and ideas are provided in Module 4 (4.1-4.3) of the GSLS GenAI Teacher Tutorials.

Level 1: no AI:

• Closed-book pop quizzes, oral defences, or in-class debates are hard to fake and even harder to outsource
• Fishbowl discussions, group work, lab-based or field-based work are live, unscripted engagement (Level 1 excludes AI specifically, not human collaboration)

Level 2: AI supports exploration before a student commits to anything:

• Coding exercises where AI helps explore possible approaches before a method is chosen
• Preliminary literature or background searches, before committing to a specific research angle

Level 3: AI reacts to work that’s already the student’s own:

• Debugging exercises, where the student has already written the code
• Writing and argumentation checks, where the student brings a completed draft and AI flags weak points. The student then decides what to do with that. The tool never rewrites it for them

One more technique that works at any level: ask students for their own conclusions on a topic, supported with specific evidence from a lecture or class discussion. For example, “Given our discussion of [topic], what conclusions can you draw, and what from the discussion supports them?” This asks the student’s own judgement, not just recall, and a generic AI answer can’t substitute for it since it wasn’t in the room. Even better is to have them do this in class with pen and paper!

Whatever the activity, at Levels 2 and 3 students should be able to explain how they used a tool, what they asked, what they kept, and what they changed. This is exactly what the disclosure statement is for (see Example Disclosure Agreements).