The impact of Language Models (such as ChatGPT) on PBL education

The introduction of Large Language Models, such as ChatGPT, raises questions on the role of education, and what kind of skills students need to learn and how we assess students.

In education, ChatGPT can be used in multiple contexts. For example, as a tool to generate complete essays, summarise articles and books, elaborate codes, and solve math problems.

Students can use the system as a critical friend providing feedback or, as part of a learning activity, critically assess its output. Still, little is known about the impact of LLMs on student-centred educational models, such as Problem-Based Learning (PBL), as implemented at Maastricht University.

The project has three main goals:

1. Gain scientific insights on the impact of LLMs (such as ChatGPT) on teaching staff and students in a PBL environment.
2. Gain insight into the way ChatGPT is/can be used within UM’s PBL environment through
   • several pilots in which ChatGPT is used during teaching, learning, and assessment activities
   • identifying best practices within the faculties
3. Provide UM with concrete evidence-informed guidelines for teaching staff and students to use LLMs (such as ChatGPT) in PBL education. These guidelines will be generic and faculty-specific (depending on which faculties will be involved in the project).

The project output can be summarised as follows:

• A review of the literature concerning Large Language Models for PBL education.
• A teacher guideline for using Large Language Models for designing intended learning outcomes, teaching and learning activities, and assessment.
• A collection of best practices for using Large Language Models in PBL education (based on existing initiatives and pilots).
• A student guideline for using Large Language Models for effective learning.
• Outline professional development activities to help teaching staff become acquainted with Large Language Models.

View the project proposal (on UMployee).