Alvaro Fernandez Quilez
Associate professor of Univesrity of Stavanger and a head of AI Laboratory of Univeristy of Stavanger (SAIL)
University of Stavanger has launched an online course to educate participants with a wide range of background about AI and its implications. Besides the challenges of educating a large heterogeneous auditorium the course has forced the developers to rethink and redesign the assessment and grading of homework when the students are not only allowed, but instructed to use AI tools to solve the problems.
Background
University of Stavanger (UiS) has launched a new fully digital (online) AI course in the autumn 2024. DAT105: AI introduction for everyone was conceived to cover the demand and need of society for an accessible course clarifying how to use generative AI on a daily basis, and the different ethical implications on the use of the technology. The course does not require nor assumes any previous knowledge nor technical competences, other than access to a computer and an understanding on navigating the internet. With this course, we aimed to provide a solid fundament to better understand the potential and limitations of AI for any type of audience with no previous or existing limited AI knowledge.
The target audience of the course is broad, as exemplified but our first experience running it. We observe groups of very heterogenous ages, from students in their 20s to students in their late 50s or early 60s. The background of the students was quite heterogenous, ranging from those that are currently enrolled in a bachelor program at our own university (UiS), those that are working and looking for a solution tailored to their flexibility needs or those that want to pursue a life-long learning and in demand of a course to catch up with the current technological trends.
The course currently consists of four modules, aimed to (1) provide an intuition and understanding of common AI concepts and capabilities of current AI systems (2) guide interested users in the adoption and use of AI, and to gain an general understanding on for what and when to use AI on a daily basis (3) showcase from a practical perspective common generative AI tools, their use and their possibilities and (4) highlight some of the current limitations, ethical debates and legal landscape surrounding AI tools. All modules have several video-lectures recorded in English and with available subtitles in Norwegian. This allows students to complete the course at their own pace. In addition, all modules include quizzes that are not formally evaluated but intend to give students a way to measure their own learning. The course is formally evaluated through a final project deliverable. The students can start working on the deliverable from day 1 of the course, providing them a total of 3 months to fulfill it.
The final project is a “hands-on” assignment, where the students are tasked with reflecting on a possible use of AI that they could find interesting. The students define their own problem, necessary and provide some reflections on how AI could help them solve that issue and why. However, the assignment has a big catch: students are forced to use at least two generate AI tools to solve it. This includes AI to generate pictures or even inspiration in the writing. Whilst the approach might seem daunting, we seek to evaluate the student ability to use AI in an ethical way and human-centered way whilst evaluating their understanding of possibilities and limitations of the technology. For that purpose, students are required to disclose their conversations with AI, and prompts used as part of it. This helps us define whether the use was as intended, or not.
Results
DAT105 started with 1123 students, from which around 1000 finished the course. Out of the 1000 students, 352 students delivered the final project assignment. Around. 90% of students passed the exam, showing high creativity to solve problems with the use of AI that were relevant for them in different everyday situations: job, studies or even home routines. After finalizing the course, the student feedback was overall positive. Most of the students (irrespectively of whether they presented themselves to the exam or not) mentioned that they felt they had a much better understanding of the AI technology and its possibilities (and challenges) after the course. This was irrespectively of the background of the student, showing that the course was able to reach a wide population. After critical reflection and analysis of the results, we see them as a very positive sign that the course was well-received and that students reached the learning goals. We correlate the rate of students that did not take the exam to the digital nature of the course. In our experience, digital courses tend to have an exam attendance rate below 20%. All things considered; we deem our results as very satisfactory.
Impact on the teacher community
DAT105 started as a pilot, since our evaluation methodology is quite different compared to standard ways of evaluating courses of the same nature. In DAT105, we assume students will use the AI technology and embrace that fact. Our evaluation is thus conceived to disentangle our desired learning outcomes from the use of the technology itself, through making its use an integral part of our deliverable. We foresee this approach as something that might become more common in the AI era; re-thinking evaluations such that account for the new available technologies and adapt to what potentially society might demand as new skills from our students. In this case, we place emphasis in the reflection capacity of students. We think that our own experience can serve to exemplify the demand for more “application-oriented” courses, and the need to adapt our teaching to the current times.
Our opinion on the course
Overall, we consider ourselves satisfied with the results and experience of our students. However, we acknowledge that no evaluation system is ideal and, in this case, we observed a considerable time-demand to properly evaluate the deliverable due to its nature (prompts, chats with AI, the report itself…). We also foresee a need to continue developing the course, as AI technology rapidly evolves, and some contents of the course might already have fallen behind even after 1 year of its launch. Our experience with DAT105 has given us some optimism on possibilities of digital education and has prompted the development of other digital courses at UiS.
Course continuation
The course is intended to be run on a yearly basis and will continue this upcoming semester (August-December 2025).
Link to the course description https://www.uis.no/en/studies/new-ai-course-its-for-everyone