Opportunities with generative AI in higher education

A central aspect in using generative AI is that it should not be used to produce fact-based material unless you yourself provide the basic factual sources for the service. It is also crucial to ensure the quality of the generated material before using it. Remember, even if it is produced with the aid of AI, you are still the author. As with all uses of generative AI, it is also very important to make ethical and legal considerations.

Examples of potential uses include:

• Developing materials for teaching activities (e.g. cases)
  In some subject areas, it might be relevant to let students work with so-called cases. Here, generative AI can be used to create multiple cases within the specific subject area, which can save time and cater for relevant teaching situations. An additional advantage is that these cases are fictional, thus avoiding the risk of affecting individual persons (e.g. using patient cases in healthcare education). However, it is very important to review the cases generated for content accuracy and how they will function in the teaching situation. If producing good cases proves to be a challenge, consider using follow-up instructions to refine or develop them further.
   
• Generating ideas for teaching plans
  When preparing teaching, generative AI can be used to come up with ideas for teaching plans. This might involve inputting the course's learning outcomes and forms of examination and asking for suggestions on a teaching plan. It is also possible to get more specific guidance on the design of certain teaching activities based on the same data. Again, it may be beneficial to ask for refinement of ideas, more precise wording, etc., as a satisfactory answer may not always be generated from the first prompt.
   
• Support in Formulating Feedback
  Some generative AI services, such as ChatGPT (paid version), are adept at mimicking a way to formulate text based on input data. Based on this, generative AI can be used to produce various types of texts, e.g., feedback. It may suffice to make support notes which are then input into the service, which reformulates them into a cohesive text in the style it has been told to imitate. Of course, teachers and/or examiners need to do the actual assessment work. As part of the role of an official, there is a responsibility to ensure that content is accurate, so it is also absolutely necessary to review what is generated to ensure it aligns with the intended feedback.
   
• Support in Text Production
  Since language models are trained to mimic human language, they can also be used in various ways for text production. This might involve summarizing texts, getting ideas for new texts or suggestions for changes, rewriting in a certain style, help with translation, and more. Here, the possibilities need to be explored based on the perceived needs.

Of course, there are a wide range of additional possible uses for generative AI in higher education. More examples of opportunities, particularly with ChatGPT, are presented in a webinar organized by PIL, where there will also be an opportunity to share personal experiences.