Good, safe and responsible use of AI
For some time, there has been a demand for training and clear guidelines for the use of AI at UiO. The AI team at the IT Department has now collaborated with the University Library (UB), the Department of Organisation and Human Resources (OPA) and the IT Department’s legal team to develop both a course and two sets of guidelines: one for the use of generative AI and one for the use of agentic AI.
There has been internal discussion in the IT Department about what the course should contain. “Everyone can use GPT UiO and other generative AI,” says the head of the AI team at the IT Department, Senior Engineer Maiken Pedersen. “You just type in a question and you get an answer. But since UiO aims for AI use to be good, safe and responsible, a bit more is required,” she explains.
Course with knowledge content and tasks for practice and reflection
Before the summer, the IT Department sent out a survey entitled "Artificial Intelligence: competence and use" to all administrative and technical staff at UiO. Maiken Pedersen highlights an important finding from the survey: “We received nearly 700 responses, and of these, 50% had not completed any form of training on AI. As a workplace, we have a responsibility to help our employees be both well informed and well prepared when facing this groundbreaking technology. The course now being launched is intended to establish a shared basic AI competence as a starting point for the work of raising AI competence among staff at UiO.”
The course is in Canvas and is available to all employees at UiO. It is at the moment available in Norwegian only. Each chapter has clear learning objectives, and the entire course takes about 1 hour to complete. Since the course is divided into chapters, it is easy to take breaks and continue later. In the course, you will learn, among other things, about the AI tools we have available at UiO and what each of them is best suited for. You will learn how to use the tools to get the best possible results, and not least, you will learn about the limitations of the technology.
As a participant in the course, you also get access to a number of tasks that show the difference between good and less good use, give you practice in prompting, and let you reflect on AI and how it affects our everyday work when we always have an oracle available. In the course, we also take a broader view and learn about manipulation of language models, model biases, and the impact of AI technology on climate and the environment.
Guidelines that safeguard UiO’s reputation and security
“I am very pleased that we are now launching both a course and guidelines for good, safe and responsible use of AI,” says IT Director Gard Thomassen. “Many people already use AI in their work, but are unsure about what is acceptable and what is not, and not least whether they are using this new technology correctly. But even those who have come the furthest and have the greatest competence need guidelines that clearly state how and in what ways AI can be allowed at UiO,” Thomassen continues. He recommends that all employees take the course and familiarize themselves with the guidelines.
The guidelines for agentic AI are only for those who use or intend to use so‑called AI agents, but the course and the general guidelines for AI use are for everyone – including those who “only” use GPT UiO. The IT Director wants there to be good conditions for sound administrative use of AI at UiO, and for researchers and teachers to be able to try out new tools, but he also has a warning.
“If we use AI incorrectly, it can affect both UiO’s reputation and the security of the data and information we develop and manage. As a respected learning and research institution, we cannot simply ask GPT and take the answer at face value. Nor can we unleash agents on personal data and research data without being completely sure we know what we are doing. That is why the guidelines and the knowledge from our course are so important,” Thomassen explains.
But the guidelines are not set in stone. “This means that the restrictions they impose are very much in force, but if anyone discovers unnecessary limitations or other things that should be changed, they must speak up. The guidelines will also be revised to keep up with developments in the field. I know that we have some of the best people in Norway – and in Europe in some areas – so we will be attentive,” concludes the IT Director.
Read the guidelines, and take the course
- UiO’s guidelines for the use of AI
- UiO’s guidelines for agentic artificial intelligence
- COURSE: Introduksjon til kunstig intelligens for ansatte ved UiO
- The course is at the moment available in Norwegian only - an English version will come later.