Beyond Answers: How LLMs Can Pursue Strategic Thinking in Education

TL;DR

This paper explores how Large Language Models can act as both patient tutors and collaborative partners to enhance education delivery and illustrates how LLMs can make education more inclusive and engaging while empowering students to reach their full potential.

Abstract

Artificial Intelligence (AI) holds transformative potential in education, enabling personalized learning, enhancing inclusivity, and encouraging creativity and curiosity. In this paper, we explore how Large Language Models (LLMs) can act as both patient tutors and collaborative partners to enhance education delivery. As tutors, LLMs personalize learning by offering step-by-step explanations and addressing individual needs, making education more inclusive for students with diverse backgrounds or abilities. As collaborators, they expand students' horizons, supporting them in tackling complex, real-world problems and co-creating innovative projects. However, to fully realize these benefits, LLMs must be leveraged not as tools for providing direct solutions but rather to guide students in developing resolving strategies and finding learning paths together. Therefore, a strong emphasis should be placed on educating students and teachers on the successful use of LLMs to ensure their effective integration into classrooms. Through practical examples and real-world case studies, this paper illustrates how LLMs can make education more inclusive and engaging while empowering students to reach their full potential.

Figures (6)

• Figure 1: Representation of the dual role of LLMs in education, as tutors and as collaborators. In both the scenarios, it is crucial that students interact with the LLM not to achieve the solution but to pursue resolving strategies and a deeper understanding of concepts.
• Figure 2: Example of interactive and collaborative usage of an LLM from a student trying to solve a classical math problem in a math recovery course. The student does not stop at the first answer giving the solution to his problem, but rather tries to understand the process and the theory behind the result. In the end, the student has learned the theory behind this problem being able to solve similar examples. The example is conducted with ChatGPT-4o openai2023gpt4.
• Figure 3: A group of students are asked to find the solution to a complex problem of which they do not have prior knowledge or background. First, the group interacts with the LLM to understand the problem and the suggested model to solve it (Step 1). They try to bring back the problem to an easier scenario they can better understand. Once understood, the group starts trying to find a strategy to solve the problem together with the LLM and the help of a CAS software (Step 2). In the end, the students find a strategy to solve the problem and obtain satisfactory results (Step 3). Chats are cut for visualization purposes. The example is conducted with ChatGPT-4o openai2023gpt4.
• Figure 4: Guided by the curiosity to understand the epidemic spread of COVID-19, the students expressed an interest in studying epidemic spread models and realizing a simulation on the data from their Italian region. Here, at first, the LLM helps students navigate the material provided by the teacher and understand how to solve differential equations with their background knowledge. Secondly, they interacted with the LLM to realize the simulation on real-world data of their interest and estimate the COVID-19 spread. Simulation results are presented in Fig. \ref{['fig:epidemic results']}.
• Figure 5: Results at different times (Day 0, 20, and 145) of the study developed in the case study: epidemic spread. Simulation on a population of 90000 people for 200 days, no vaccinated people at the beginning, and vax life of 50 days. Simulated 4 interactions per person per day. After just 20 days, the epidemic already significantly spread infecting a considerable number of people. After 145 days, the people started to get vaccinated and the epidemic spread crucially reduced.