Learning outcome
The goal of this course is to provide insight and experience on how language models such as ChatGPT work. The course is designed as a digital workshop (2 hours) with a focus on testing ChatGPT on your own exam questions. The topics we will cover are:
- What is a language model and how does it work? What tools are we talking about?
- What should I consider when creating exam questions?
- What is important to include in an assessment guide?
- Can a human assessor recognize text written by AI?
- What should I do if I suspect cheating?
The course is particularly aimed at course conveners and examiners in courses with take-home exams/term papers, but is open to other.
A Zoom link will be sent to all registered participants prior to the start of the workshop.
Preparations
- Read and assess a student paper
- Which grade would you give it?
- Do you suspect that the student have used AI-tools? Please highlight the parts were you suspect auto-generated text.
- You should bring 1-2 previous/planned exam questions.
- If you don't have any, we will provide you with one
- Suggested reading:
Agenda
12:00-14:00
- Welcome: Introduction and clarification of learning outome
- What is a language model? A brief introduction with examples and comments from participants
BREAK
- Case 1: Testing the language model on your own exam question with discussion
- Case 2: Testing the language model on an exam response with discussion Summary/reflections
14:00-15:00
- Optional meeting hour: here you can further test the tool, ask questions, and get help.