Project

A pilot study on training chatbots for administrative and technical student support in courses

Code
EXT/ONZ/000231
Duration
01 May 2024 → Ongoing
Research disciplines
  • Natural sciences
    • Information technologies
  • Social sciences
    • Health, education and welfare economics
    • Higher education
Keywords
chatbot student support automation artificial intelligence
 
Project description

In recent years, and particularly in 2023, major advances in artificial intelligence have opened completely new possibilities to improve learning and teaching experiences. New technologies and solutions are constantly showing up and rapidly maturing. Such innovations align well with the new generation of students who are raised in a fully digital environment. Students demand and need information at a faster speed than instructors can sometimes provide. Often, information is already online, but looking after it, particularly during stressful situations such as before the exams and other assignments, is just too tiring. More interactive and speedy solutions are necessary. This project aims to address this problem by piloting the use of chatbots to assist students with the administrative and technical content of courses. A chatbot is essentially a computer program that simulates and processes human conversation, allowing humans to interact with digital devices as if they were real people. Online shops, banks, and some government agencies have been using such tools for customer service. The available technologies are mature enough to be used in the academic context to support students and instructions. It is, however, fundamental to be cautious with artificial intelligence. For this pilot experiment, the plan is to first train a chatbot with administrative content (e.g. deadlines, course/assignment rules, class schedule, materials) and then add some introductory technical concepts (e.g. basic definitions) of the same course for entertainment and engagement. The project aims to evaluate two specific solutions in the market. The first is more creative and engaging by design, and thus able to provide general support to students (e.g. motivation, study and organisational tips, suggestions) using online information. The second is more formal and business-oriented (i.e. less engaging and designed for customer support) but able to return specific information as added to the training database. With those available tools, it is possible to define a digital persona (a.k.a. avatar) and train different chatbots for different courses, also locking student feedback to avoid malicious training and use. We will train chatbots in two courses of our programme in Economics to be later deployed and tested with students in real-life conditions. We foresee enormous potential for such tools across the faculty and beyond. If successful and reliable, this solution could be independently implemented in all courses in our programmes, being particularly interesting for large introductory courses (300-400 students) to reduce the burden on instructors. We will estimate the necessary training and testing time (i.e. labour cost), evaluate potential challenges and limitations, and interpret the student feedback, to inform colleagues on the cost-benefit to implement this solution on their own courses. We finally plan to provide a tutorial for colleagues who wish to implement the tool themselves