Can artificial intelligence be useful in professional education?

This article has been translated from Norwegian using artificial intelligence.

We are responsible for each of our courses in the master's program in social work, where approximately 30 students attend three gatherings during the semester. The courses are called Innovation and Service Development in Partnership (10 credits) and Worldview, Values, and Relationships in Professional Practice (VID's profile course of 5 credits). It is mostly the same students who take both courses.

When we were going to experiment with AI in the courses, we chose to collaborate on a joint program across the courses. This consisted of a common introduction to AI and common reflections during the semester. This was intended to support separate assignments in each of the courses.

The purpose of trying out AI in this way has been to promote awareness and build experience with the use of AI in teaching and professional practice.

In the innovation course, students have used a chatbot in part of a larger project work (the project work is the course requirement). In the profile course, students have generated text with AI, which they reproduced and analyzed in their exam papers (home exam).

The purpose of trying out AI in this way has been to promote awareness and build experience with the use of AI in teaching and professional practice. In both courses, reflection on how AI can be used constructively in studies, projects, and professional practice was a central part of the student work.

Assignments

In the innovation course, students have a project work where they work in groups throughout the semester. Here, students were asked to develop and ask questions (prompts) to a chatbot (GPT UiO) that were relevant to their project work. They were also asked to reflect on how they could assess the quality of the answers they received, as well as how the answers could be used in the further project work. At the subsequent gathering, we conducted a workshop across the courses with reflection on this task. We also challenged a more general reflection on how a chatbot can be used in professional practice in social work.

In the profile course, the exam assignment was for students to generate a short text with AI on a topic in the course. They could decide for themselves how to formulate their prompt. They were then to compare the academic content of the AI-generated text with self-selected curriculum texts and reflect on whether the text they generated with AI provided relevant input to professional practice in social work in Norway today. Students were offered support to generate text with GPT UiO on the same day the exam assignment was distributed.

Students were asked for consent to use group submissions and exam papers in research and development projects on AI in social science education.

Experiences and reflections

Feedback from students about AI in the group work varied from skepticism to enthusiasm. Throughout, they reflected that it requires professional knowledge, understanding, as well as ethical and critical reflection to assess the answers the chatbot gave. They said it was difficult to assess the answers they had received early in the course (semester). When the project work was completed, they reported to a small extent that they had used the AI-generated material further in the group work. We wonder if this may be related to the fact that they have worked further with the material after the AI assignment, and that they eventually experienced the material as their own (rather than as AI-generated material).

In the oral evaluation we conducted in the last gathering of the semester, students were generally positive that we were experimenting with AI in the courses. One student said that it was "Positive with a focus on the use of AI as a tool. Positive that you are not afraid to engage with it." The feedback from the students also gave some thoughts on what we could have done differently.

It varied how well the students felt the introduction to AI hit them. In the oral evaluation, one student thought that "the teaching was 'a bit in it'," while another wondered if the teaching could have been optional since it had not been so useful for students with more knowledge of and experience with AI. In retrospect, we think that we could have used this range as a resource in the teaching with a well-planned group work. If the students had worked in mixed groups, those with more knowledge of and experience with AI could have guided the others, while the students with less knowledge of and experience with AI could have challenged fellow students with questions and reflection.

...it may be interesting to give students more detailed instructions on how to write good prompts, either in the teaching or in the assignment text.

During the grading of the exam papers where the students had used AI, we noticed that the content in the texts the students had generated with AI in most cases was very general and high-level. Several of the students commented on it. "AI 'lists' the concepts in the text, without further elaboration, interpretation or problematization," wrote one student. Students who had written more specific prompts received more specific answers. In light of this, we think that it may be interesting to give students more detailed instructions on how to write good prompts, either in the teaching or in the assignment text.

Conclusion

We wanted to have a professional focus on AI and that technical issues should not take up a disproportionate amount of space, neither in the teaching nor in the courses. We felt that this was reflected in the students' reflections and evaluations, and that the students treated AI as a tool on par with other tools in the project work.

The semester has taught us that AI becomes more interesting in professional contexts when you use well-developed and good prompts. Perhaps further experiments with AI in professional education should address how to work with writing prompts as a professional question?

Read more?

Read the chronicle in Khrono.