Learning cannot be outsourced to AI – This is how AI is transforming university education

Learning cannot be outsourced to AI - This is how AI is transforming university education

In recent years, advances in AI, particularly generative AI applications such as Copilot and ChatGPT, have become increasingly dominant, including at the University. We asked Professors Kalle Juuti and Anette Alén what role AI plays in their teaching.

Responsible AI skills

Both Juuti and Alén have introduced generative AI applications in their teaching, as they wish to teach their students responsible AI skills. Alén uses the University's tool, for example, in her first-year group teaching. On the course, students compare a legal case and CurreChat's responses, considering the use of AI in legal information retrieval.

In an advanced course Juuti teaches, students complete a demanding development project, or a solution concept. During the course, they are encouraged to use CurreChat at certain stages, including for brainstorming support and feedback. At the same time, students gain experiences of using generative AI at different stages of extensive assignments.

AI changes teaching

However, generative AI is not suited to all university studies.

Professor Juuti emphasises that when introducing a new technology such as AI, you should consider how it changes teaching and studying. This means getting back to basics: the goal of teaching and how to assess the achievement of this goal.

For a basic course in education he teaches, Juuti has switched to essays completed in person in the lecture room as a means of assessing students.

"An essay is an effective way of demonstrating that students are able to argue on certain fundamentals of the field. When students write essays in the lecture room, the teacher need not wonder about the origin of the ideas expressed in the text. Having said that, assessment always depends on the course learning outcomes," says Juuti.

Professor Alén too points out that circumstances where students' skills are to be measured in a controlled manner, for example, by organising invigilated examination sessions, should be identified at the degree level. It is equally important to identify situations where the use of AI tools cannot be controlled, and to strive to guide and set limits for such use.

Transparency and reflection

E-learning Specialist Sanna-Katja Parikka recommends that the use of generative AI be openly described already in the course description.

"Such an AI clause should describe, for instance, how students are and are not allowed to use these tools throughout the course and its assignments."

Parikka also recommends that any course assignments involving generative AI tools should include a critical reflection component guiding students to assess AI use and the content it generates.

"The purpose of reflection is to deepen learning, support critical thinking skills and verify the learning process. In the best case, the critical reflection section is strongly linked to assignment- and course-specific learning outcomes."

Appropriate guidelines and tools

Both Juuti and Alén commend the University of Helsinki for providing a good starting point for considering AI in teaching. In 2023, the University established its first guidelines for the . Among other things, they prohibit the use of language models in maturity tests.

According to Alén, who is also one of the vice-deans of the Faculty of Law, the Faculty has drawn up supplementary guidelines pertaining especially to theses. While generative AI can be used to support writing for master's theses, students are required to describe transparently how they do so and to contemplate AI use beforehand.

CurreChat also receives praise from both Juuti and Alén.

"It's great that students can openly use generative AI with their own University credentials as part of their studies instead of resorting to dubious free versions with questionable information security," Juuti praises.

Participate in training and experiment

For colleagues wishing to develop their pedagogical skills in relation to AI, Alén recommends University training offerings. She also urges them to try out CurreChat on a familiar topic, making it possible to meaningfully assess whether AI produces sensible answers.

According to Parikka, the University's Educational Technology Services unit offers commissioned training in CurreChat use for teachers.

In her course, Alén piloted an AI chatbot representing the client with whom students conduct a legal discussion at the beginning and end of the course.

"Such a task develops not only skills in legal counselling, but also AI literacy. Having the same discussion at the beginning and end of the course also demonstrates your learning", Alén says.

Professor Juuti emphasises that taking AI into consideration requires, particularly within disciplines, co-creation and contemplating areas in which AI should be used and areas where skills must be demonstrated in other ways. Such co-creation takes up teachers' working hours and requires the opportunity to focus on teaching development.