These guidelines are being tested. This means that we welcome feedback on whether anything should be changed to make the guidelines more comprehensive, easier to understand, or otherwise improved. Contact us at ai-contact@usit.uio.no.
Main rule: Use approved services in your work at UiO. This means that you may not use non-approved AI services to process UiO data. This is particularly important for data classified as yellow, red, or black.
Always follow internal guidelines for your AI tool, where such guidelines exist.
Choose the right tool. You must use AI tools that are approved for the protection level required by your data. See how to choose the correct protection level.
Be especially careful with personal data. Even if you have chosen the right tool for the protection level required by your data, you cannot use the data in any way you wish. Laws and regulations concerning personal data also apply here.
You must be able to account for, and are responsible for, all decisions made by the AI if they affect the organisation. In cases where AI makes decisions on behalf of the organisation, these decisions must be transparent and verifiable. AI should not make decisions in case management (NO: saksbehandling), but may be used to support decision-making.
Always quality-assure AI-generated content. Always review what the AI produces, and never use text or code that you do not understand.
Do not use AI uncritically as a reference source; there is no guarantee that what is presented as fact is true.
Disclose the use of AI. If you include AI-generated content in work that you submit or deliver—for example, in assignments, publications, text intended for the public, or official documents—state that you have used AI, how you used it, and which AI services were used.
Be aware of the rights associated with data in prompts and uploads. The material may be subject to restrictions on lawful use in AI tools, for example relating to copyright. You must always clarify the rights to the data/the material you intend to use before doing so. Data includes both instructions (prompts) and files that are uploaded or added.
Be aware of your own responsibility in research. Researchers themselves are responsible for ensuring that the use of AI has a legal basis, that methods and results are reproducible and accurate, and that the research complies with applicable ethical and legal standards.
UiO emphasizes the responsible and sustainable use of AI. This means that you must be aware of the environmental and ethical aspects of the technology. In some cases, this may mean refraining from using AI, limiting its use, or choosing an AI system with a smaller environmental or ethical footprint.
Seek advice. When using AI, situations may arise where you are uncertain about data classification, ethics, or methodology. If in doubt, speak to your leader or supervisor.
Notify us if you discover or suspect that AI-use has led to a security breach. Contact UiO’s IT security team at cert@uio.no. They can help you handle the matter.
This always applies
All use of AI on UiO equipment and/or with UiO data must comply with the provisions in the guidelines on this page.
In addition, UiO’s other internal guidelines, as well as applicable laws and regulations, apply. This includes, but is not limited to:
- UiO’s information security management system, including the classification guide and storage guide
- Data protection regulations and UiO’s guidelines for processing personal data
- Academic integrity and research ethics
- Rules governing administrative case processing
Specific guidelines for agentic AI
If you intend to use AI agents—that is, AI that can perform actions on its own initiative or on your behalf—an additional set of guidelines applies. You are responsible for ensuring that such use takes place in a controlled, secure, and responsible manner.
See also
How to use AI as a student: guidelines, recommended tools and useful resources