Norwegian version of this page

Artificial intelligence in teaching and assessment – what should you consider?

At the initiative of the education leadership meeting, the Vice-Dean for Studies established a working group on AI in teaching and assessment in December 2024. Here you will find both the AI group’s recommendations and a summary of the other topics discussed.

Banner

Photo: Jarli & Jordan/UiO

Examples from teachers at the Faculty of Social Sciences

Øystein Sørensen - PSI

 

Signy Irene Vabo - ISV

Understanding AI and the possibilities offered by these tools

To use AI in a productive and responsible way, it is important to understand the technology behind these tools, as well as the possibilities and limitations they entail. This means that you should familiarise yourself with ethics and social responsibility and legal perspectives before you or your students use such tools. Also see the Guidelines for AI-generated content in UiO’s digital channels, and familiarise yourself with UiO’s AI information pages for both teachers and students.

What is generative artificial intelligence?

GPT uses large language models (LLMs), a type of machine learning focused on processing and generating human-like text. These models are designed to recognise language patterns and structures based on the vast datasets on which they are trained. Because of their size and complexity, they are exceptionally good at understanding context and identifying meaningful connections. It is important to recognise that language models do not function as knowledge databases. Although they can be used effectively in many different areas with appropriate guidance, it is essential not to assume automatically that the response you receive is correct or academically sound. Under “Recommended resources”, you will find several resources that can help you understand and explore the technology further.

Recommended resources

The resources and descriptions presented in this section are taken from UiO’s central information page for teachers.

Understand the technology

  • The video Introduction to AI for Teachers and Students, produced by the Wharton School. The video explains what artificial intelligence is. It demonstrates AI tools other than GPT UiO, but these are based on the same language technology. Running time: 10 minutes. Available in English.
  • The video Large Language Models (LLMs), produced by the Wharton School. The video explains large language models. It demonstrates AI tools other than GPT UiO, but these are based on the same language technology. Running time: 12 minutes. Available in English.
  • The guide “ChatGPT and Artificial Intelligence in Higher Education: Quick Start Guide”, published by UNESCO. It provides a short introduction to AI, examples of its use and an outline of key critical issues to consider. Estimated reading time: 30 minutes. Available in English.
  • The webpage What are AI tools and how do they work?, created by the University of Groningen. It provides a brief explanation of how AI tools work and the types of AI models that exist. Estimated reading time: 5 minutes. Available in English.
  • Harvard metaLAB’s AI Guide is an interactive guide that helps beginners become familiar with AI and explore how it can be used in teaching. Available in English.
  • The book Maskiner som tenker – algoritmenes hemmeligheter og veien til kunstig intelligens [Thinking Machines: The Secrets of Algorithms and the Path to Artificial Intelligence] by Inga Strümke. It explains artificial intelligence thoroughly and accessibly, and provides insight into a range of AI-related opportunities and challenges. Available in Norwegian.

Explore prompting

  • The video Prompting AI, produced by the Wharton School. The video shows how to ask useful questions. It demonstrates AI tools other than GPT UiO, but you can use the same approach. Running time: 11 minutes. Available in English.
  • The video AI for Students, produced by the Wharton School. It gives teachers useful insights and inspiration for learning activities and prompting ideas. Running time: 10 minutes. Available in English.
  • Guides for GPT UiO. The IT Department’s webpages provide technical guidance on how to get started with GPT UiO.
  • The website of KURT – UiO’s Centre for Teaching in Science and Technology – includes tips and examples of instruction texts that you can add to GPT UiO.

Stay up to date

Websites
  • Futuretools provides an overview of many of the AI tools currently available.
  • AI Magazine is regularly updated with the latest news on AI developments.
YouTube channels
  • Matt Wolfe provides weekly video updates, demonstrations and potential use cases for new tools.
  • AI Explained provides regular updates on major trends in AI development and concise summaries.

Basic rules for using AI at UiO

As a teacher, you are responsible for deciding whether, and if so how, AI may be used in a course. Students are encouraged to contact you if they are unsure about what is permitted. The rules for using AI may therefore vary between subjects and assignments.

Rules communicated to students

These are the basic rules for using AI presented on the student information page:

  • Use AI as support, not as a replacement for your own work.
  • Clearly indicate when and how you have used AI.
  • AI can be a useful supplement for understanding topics, course material and terminology, but always verify information against reliable sources.
  • Avoid entering sensitive information and copyrighted material into AI tools without approval.

 

Approved AI tools at UiO

UiO has four approved AI tools:

Course planning

Whether or not you choose to use AI, you should actively consider the implications of this technology for your discipline. Reflect on the knowledge and skills students need to meet future demands in working life and society, and consider how these implications are reflected in, and addressed through, the intended learning outcomes, assessment methods and learning activities in your course.

What should you consider when planning your course?

Course design

Regardless of whether you choose to use AI, it is essential to design learning activities that are closely aligned with the course’s intended learning outcomes and the way students are assessed.

When planning a course, it is important to reflect on how AI can strengthen learning, when and how it can be used, and whether AI should be avoided in specific learning and assessment activities in order to safeguard key learning processes. You should also consider ethical issues and how students can demonstrate academic integrity if they use AI. In addition, it is recommended that you reflect on the demands of working life and society, what students should learn, and how this is addressed in your course.

Remember that you are students’ main point of contact. Make a plan for how you will keep up to date with developments in the technology, and ensure clear communication with students on the subject. Specify on the course page whether, and how, the use of AI is permitted, and make sure that any recommended use of AI does not conflict with the privacy regulations and the GDPR.

The Faculty has developed a suggested appendix to grading guidelines. You may use parts of, or the entire, document in your grading guidelines.

Learning activities

In higher education, reading original sources, producing one’s own texts, and independent exploration and problem-solving are essential to developing students’ capacity for learning, reflection and critical thinking. There may therefore be good reasons to include such activities in ways that are fully or partly shielded from AI, but there may also be good reasons to allow the use of AI in these activities.

If you wish to include AI in learning activities, it is nevertheless recommended that this is done in a way that strengthens students’ learning and metacognition, as well as their understanding of the technology and how it can be used appropriately in learning. It is further recommended that learning activities are designed as an iterative process, in which students document any dialogue with AI and justify the steps and assessments they make. Students should also be encouraged to work together to discuss and practise evaluating AI-generated responses.

Under “Learning activities with AI?” on UiO’s central information page for teachers, you will find examples of roles AI can play in teaching, as well as suggestions for learning activities with and without AI.

Assessment activities

When it comes to assessment activities, there may be good reasons to permit the use of AI in work that is assessed as well. However, you must be aware that this can make it more difficult to assess what students have actually learned.

There are no reliable plagiarism detection tools that can identify AI-generated content and are permitted for use at UiO. It is therefore recommended to focus on developing assessment methods in which the primary emphasis is on the process of learning, rather than solely on the final product. UiO’s Regulations for Studies and Examinations allow for a wide range of assessment methods. In some cases, however, a written on-campus exam will be necessary, either during a transitional period or as a continuing form of examination based on the selected learning outcomes.

Under “Assessment activities” on UiO’s central information page for teachers, you will find advice on what to consider if you permit the use of AI, as well as suggestions for assessment activities.

Recommended resources

The resources and descriptions presented in this section are taken from UiO’s central information page for teachers.

Grading

Various online tools claim to detect the likelihood that a text was generated by AI. These tools cannot be used to check whether students’ examination submissions have been wholly or partly generated by AI. UiO does not have a data processing agreement with these tools, and their use would conflict with the GDPR.

Methods for distinguishing between human-written and AI-generated text

  • Familiarise yourself with how a chatbot, such as ChatGPT, responds to and solves tasks.
  • ChatGPT itself suggests looking for the following indicators:
    • Unnatural language. Unnatural sentence structure and unusual word choices may indicate AI-generated text. Chatbots do not have the same linguistic intuition as humans and often lack variety and creativity in their use of language.
    • Repetitive language. AI models often generate text that repeats itself and lacks variation.
    • Nuance and detail. AI tools may produce text that generalises information and fails to account for nuances in language. They may also be poor at giving detailed answers to a question. Chatbots also do not have personal opinions.
    • Grammatical errors. Look for unusual word choices or grammatical errors.
    • Lack of contextualisation. AI-generated text may lack coherence or context. The text may therefore appear vague and unclear.

Is the use of AI in take-home exams and coursework cheating?

  • The regulations on cheating and the use of sources apply to both exams and compulsory assignments, with the same possible sanctions and evidentiary requirements.
  • It is not permitted to generate all or part of an examination submission using AI tools, even where the course description states that all aids are permitted. Submissions must be students’ own work.
  • If an examination assignment provides for the use of AI, the assignment text must explicitly describe how it may be used.
  • The ordinary procedure for suspected cheating applies. Contact the student adviser as soon as a suspicion arises.
  • Is the use of GPT UiO cheating? KURT (the Faculty of Mathematics and Natural Sciences) has conducted an investigation: Would GPT UiO have had to resign as Minister of Education?

Examples from teachers at the Faculty of Social Sciences

Lars Erik Kjekshus - ISS

Esten Høyland Leonardsen – PSI

 

Points of contact

This resource was developed at the request of EILIN’s learning contact network.