Generative AI in teaching and assessment at NHH
Guidelines for staff and students.
Final version after discussions in Programledermøte, Utdanningsutvalget, Rektors ledergruppe and Felles ledergruppe, December 2023.
Generative AI systems, in particular those based on Large Language Models, have recently become more powerful and more accessible, and their use has become more widespread. ChatGPT (from OpenAI) is one example, but generative AI is also increasingly incorporated in existing software. This development provides us with new opportunities and challenges. At NHH we want to explore what the technology can do and learn how to make use of it in our study programmes. At the same time, we want our students to have the skills necessary to understand the legal and ethical implications of the new technologies.
As a starting point, we would like to emphasise:
- That our main priority is to ensure that the students meet the learning outcomes of the course/programme;
- That we require assignments to contain students’ own original work;
- That generative AI is potentially useful in many of the subject fields taught at NHH, and can also improve the learning process of the students;
- That there is a need to acknowledge the use of generative AI where it is (allowed to be) used;
- That AI software should be equally accessible to all students, to ensure fair evaluation;
- That the manner in which AI software is used, should take into account ethical and legal considerations.
With regard to both teaching and research, the various disciplines and fields of study are affected differently by the use of Generative AI. Consequently, it will not be useful to have a general, overarching NHH AI-policy, that pertains to all courses and all types of assignments and assessments. While we make some recommendations, it is up to each course responsible to clearly inform the students of what the specific regulation for their course is, both in general and in the case of specific assignments. Some assignments may explicitly ask the students to work with AI tools and to analyse and evaluate the content generated, others may specify that AI tools should not be used at all, or only used in specific ways. The requirements should be clearly stated in connection with all assignments.
Recommendations for staff
Ensure that the learning activities enable the students to meet the learning outcomes of the course, and that their use of generative AI does not impede your ability to assess this. Consider alternative ways of assessing student work that ensure that you can confirm that the learning goals have been met (for instance oral exams, school exams or presentations).
Provide clear and unambiguous rules about the permitted use of AI in general, as well as for specific assignments. These rules should be easily accessible in Canvas, in a document labelled “Generative AI regulations – [Course code]”. One helpful way to implement this, could be a form, listing all assignments, their deadlines and their individual generative AI rules, as well as the manner in which the assignments are related to the learning outcome of the course.
Emphasise the expectation that the assignments should contain the students’ original work, and that there should be an acknowledgement of the use/non-use of generative AI. Students may also provide a “generative AI statement” at the beginning of each assignment, where they acknowledge that they know the rules for the specific assignment and present how/whether generative AI has been used. This could either be a “standard” text for each assignment, or a more “reflective” note.
When AI is permitted, ensure that the students are aware of the scientific limitations, the ethical considerations and the legal regulations concerning such models, related both to input and output.
For the course responsible and other internal or external examiners, sharing students’ work with external AI tools (for instance in connection with grading/assessment) is not allowed, unless these are solutions that guarantee that the legal considerations (such as anonymity and privacy issues) are met.
Recommendations for students
Generative AI is a powerful tool, that may be a very useful tool in a learning process. However, generative AI should be used in a responsible, legal and ethically sound manner.
Generative AI can give you erroneous or fake information. It may provide you with text that you are not allowed to use, and that can lead to plagiarism problems. It may also provide unbalanced, unscientific or incomplete analysis. Engage critically with the AI tools, and remember that you are responsible for everything that you hand in to the NHH, but also everything that you feed into the AI tools.
To avoid problems: Ensure that you know and follow the specific regulations of generative AI use for the assignment or task in question. For specific courses, you will find these in the Canvas form “Generative AI regulations – [course code]”.
If you are in doubt, ask the course responsible or your supervisor.