This is a template for course-level GenAI policies in higher education. The key idea is to walk through each course activity from a student perspective, tie it to the relevant learning outcomes (and later professional work life, where applicable), and derive concrete "Dos" and "Dont's" for AI use. Such policies are not limited to assessments but cover all learning activities because their point is not to be enforceable by examiners but to be actionable for students. They should be the start of a conversation between students and teachers not the end, and they should be refined with every course instance.
📑 Template for course-level GenAI policies
There are a number of alternatives that I considered for how to deal with students using GenAI in my courses:
- Ignore AI: Focussing on the learning outcomes should be enough, right? Could be, but relying on catching (and failing) phonies and frauds through oral follow-ups reduces teaching to gatekeeping the degree, and I don't like that.
- Ban AI: A complete, unconditional ban of all AI use to deter students might have the opposite effect ("forbidden fruit") and is not enforceable anyway.
- Allow AI: Allowing all AI use by default is also naïve, as it effectively abandons our statutory mandate under the Swedish Higher Education Act to develop our students' abilities to "make independent and critical assessments" and "solve problems autonomously".
- Appeal to Ideals: Allowing AI under the condition that it "must not replace human judgement" or "affect academic integrity" sounds like it's hard to disagree with, but it's utterly meaningless for students who, by their nature of being students, have not internalized yet what either of those concepts mean.
- Adapt the Learning Outcomes: Those should be regularly reassessed anyhow, but at least for my courses, I am content with most learning outcomes as they are. If anything, I would rather add some pertaining to AI Literacy.
- Change the Modes of Assessments: Indeed, the moment when producing a convincing-looking written report no longer required deep thinking and became cheaper than properly reviewing it, that way of assessing students was outdated. But (a) for many learning outcomes, it was a questionable fit even before ChatGPT, and (b) "fixing" assessments does not do any good for the 95% of the course that come before it: The teaching and learning.
Consequence: There can be university-wide high-level statements for teachers, but AI policies for students must be actionable and directly related to whatever activities they are expected to go through during the course -- assessed or otherwise.
GenAI Policy Template, Franz Zieris, Blekinge Insitute of Technology, 2026. Available at: https://github.com/bth-dipt-teaching/genai-policy/blob/main/template.md Licensed under CC BY 4.0.
Need some inspiration for your course? Here is a collection of course policies for GenAI use at the Blekinge Institute of Technology, Software Engineering Department:
- PA1481: Digital Ethics and Sustainable Development (VT2026)
- PA1483: Software Architecture (HT2026) (earlier version: HT2025)
- PA1453: Software Architectures and Quality (HT2026)
- PA1489: Basic Software Engineering (HT2026)
- PA2534: Master Thesis Course (VT2026)
- PA2577: Applied Cloud Computing and Big Data (HT2026)
You want more background information?