AI and Your Degree: What You Need to Know
AI tools like ChatGPT, GitHub Copilot, and image generators are already part of professional software and game development. Our programme has a strategy for how AI fits into your education. This page explains it in plain terms, so you know what to expect, what’s allowed, and what you’re actually here to learn.
The Short Version
- You’re expected to use AI tools, they’re part of the job you’re training for. Using AI well is not trivial.
- You still have to prove you understand the work. Oral exams, live coding, and portfolios exist to check that you (not just the AI) can do it.
- Knowing what to build, why, and whether an AI’s output is actually good is a key skill. That’s the part AI can’t do for you.
- Rules differ by course and by task. Always check the specific AI policy for each course and assignment, don’t assume what worked in one course applies in another.
- The SDU rules regarding AI are always in place; if you have doubts, ask your teacher.
What Do We Mean by "AI" Here?
Mostly this means generative AI — tools like ChatGPT, Copilot, Midjourney, and similar systems — from the point of view of you as a student using them to learn and work. It also touches on how teachers use AI to build course material, and how AI is used professionally in the games and software industry (since that shapes what skills you need).
The Tier System: How Much AI Can I Use?
Not every course or assignment treats AI the same way. We have introduced a tier system to explain how much you’re expected to use AI. Each course will tell you which tier applies. Keep in mind that (unless it’s specified in the course) using AI is not always a requirement, but the tier expresses what use of AI is permissible in the course.
| Tier | What it means | Example |
|---|---|---|
| Tier 0 — No AI | You work completely unassisted. This proves you can do it yourself. | In-class exams, foundational coding tests |
| Tier 1 — AI as reference | You can look things up with AI, like using a textbook, but you write everything yourself. | Early programming exercises |
| Tier 2 — AI as co-pilot | AI can help generate code or ideas, but you must understand, modify, and explain everything you use. A reflection log is usually required. | Mid-programme project work |
| Tier 3 — AI as team member | AI is a full production tool. You act as the director/architect and are assessed on your decisions, specifications, and quality control. | Advanced projects, thesis work |
In short: the more advanced the course, the more AI you’re expected/allowed to use, but also the more you’re expected to critically evaluate and own the result.
How This Changes Assessment
Because AI can now write code, generate designs, and even draft essays, written submissions alone can’t prove you’ve learned something. So assessment is shifting to include more of:
- Oral exams and defences
- Live coding and live design sessions
- Process documentation: logs, commit histories, and decision journals showing how you got to your final result, not just the result itself.
- Portfolios: a body of work built up over time, showing growth.
This doesn’t mean everything becomes an oral exam. It means written work will rarely be the only evidence used to grade you.
What Skills Actually Matter Now
AI is good at producing code, art, and drafts quickly. It’s not good at deciding what should be built or why. So the programme is putting more weight on:
- Reading and evaluating AI output: spotting bugs, security issues, or bad design choices in something AI generated.
- Systems thinking: understanding how different parts of a project fit together, and how to design complex systems.
- Problem formulation: being able to clearly define what you’re building and why, before anything gets written. A good prompt strategy can make the difference between good/bad AI output.
- Domain expertise: game design theory, player psychology, and interaction design are hard to automate and hard to fake.
- Fundamentals still matter: you still need to understand data structures, algorithms, and core programming concepts. If you don’t, you can’t tell when AI gets something subtly wrong.
Why Bother With Any of This?
- You graduate with real, defensible skills, not just AI-assisted output you can’t explain.
- You’re prepared for a job market where using AI well is already expected.
- Your credential stays meaningful because assessment can’t be gamed by just prompting a chatbot.
