How do we humans actually learn best? Research shows that the best learning situation is one-to-one teaching, but this is not easy to offer at a university with over 14,000 students and nowhere near as many teachers.
However, Postdoc Fiammetta Caccavale from DTU Chemical Engineering may have an alternative to one-to-one teaching that can make it much easier to learn new things when students start a new course, for example. An alternative that she will investigate in a three-year research project for which she has received a grant of DKK 2.88 million from the Villum Foundation.
"I will spend the three years investigating the pedagogical potential of artificial intelligence, because everything indicates that there is great, untapped potential," she explains and continues:
"Many of the students already use artificial intelligence when they study, but there is rarely any systematic approach to it, and I want to change that with my project. I want to create some digital tools that make the students more skilled and at the same time help the lecturers."
AI as a mentor
An example of such a tool could be a specially designed chatbot that has been prompted with the syllabus from the course the students are taking. The idea is that students can spar with it in the same way they would spar with a mentor.
"It would be fantastic if I could develop a chatbot that could be used across courses and thus follow students throughout their studies. That is clearly the dream scenario," says Fiammetta Caccavale.
At the same time, such a chatbot would also be a helping hand for lecturers, because they could use it as a permanent digital tutor on their courses.
Across national borders
Fiammetta Caccavale's three-year project is a collaboration between DTU and Imperial College London. During the first two years, she will be attached to several courses at Imperial College London and will develop a range of AI models aligned with the courses' learning objectives.
The knowledge she gains from this research will be used to develop an educational framework for studying the use of AI in higher education. This framework will be used to demonstrate a causal link between the introduction of digital AI learning tools and improvements in students' learning processes.
"I hope to get a lot of feedback from all the students at Imperial College London who are taking the courses I will be collecting data from, because students are the ones my future digital tools are intended for," explains the Postdoc.
