Guidelines on the use of generative AI in teaching and assessment

Introduction and purpose

Technological developments in generative AI[1] are fast-moving and knowledge about and use of the technology need to keep pace with this. Generative AI affects higher education at many levels. In the light of generative AI, the University’s educational and cultural mission needs to be safeguarded but also developed. Uppsala University calls on teaching staff and students[2] to explore the opportunities and challenges brought by the technology together, in a positive spirit.

One fundamental prerequisite for the guidelines is that they need to be formulated in relatively general terms. This is because generative AI is evolving very quickly, while the legal situation regarding the technology is unclear. Moreover, the conditions for dealing with generative AI vary between educational programmes and consequently the guidelines can only address matters that are applicable to all education offered at the University. Had they been more specific, the guidelines would have been unnecessarily restrictive.

The guidelines are intended for members of the teaching staff and students, including doctoral students, and provide an overarching framework for reasonable, recommended and permitted use of generative AI in teaching at all levels at Uppsala University. One prerequisite for permitted usage is compliance with the Artificial Intelligence Act, the General Data Protection Regulation and other relevant legislation. As the guidelines make clear, they can be supplemented by local provisions[3], which may be more detailed and/or restrictive.

These guidelines are supplemented on the website by guidance and support material for teaching staff and students.

Guidelines

1. If the use of generative AI is restricted in teaching situations and/or assessments, this must be clearly indicated in written information. The restriction must be justifiable with respect to the intended learning outcomes, the nature of the task or educational considerations. In teaching situations and assessments, it must also be clearly indicated in written information if the use of generative AI must be reported.
2. Students planning to use generative AI in connection with teaching situations or assessments are responsible for keeping themselves informed of both central and, where applicable, local guidelines.
3. If students are expected to use generative AI in teaching situations or assessments, course coordinators/examiners must be able to ensure that tools are made available at no cost to the student. If doctoral students are expected to use generative AI, the principal supervisor is responsible for ensuring that suitable tools are available at no cost to the doctoral student.
4. A person who has made use of AI-generated material is responsible for how it is used. Usage must be guided by academic probity, a critical attitude to the reliability of the material and ethical considerations.
5. Sensitive personal data may only be transferred to generative AI systems approved by the University.
6. Material produced by students may not be transferred without the students’ consent to generative AI systems that might use the material to train the systems.

[1] In the following, the term ‘generative AI’ refers to systems that, in response to instructions from the user, and based on large quantities of training data entered into the system, can create new material in the form of text, images, videos, codes, music, etc. Some examples of such systems are ChatGPT, Dall.E, Gemini, Perplexity, Microsoft Copilot and Midjourney.

[2] In this document, the word ‘students’ refers to students at all educational levels, including doctoral students.

[3] For example, at programme level, in a course syllabus, before an assessment or in a doctoral student study plan.