Skip to main content Skip to navigation

AI and Education: An Investigation into the Use of ChatGPT for Systems Thinking

Project Overview

The document examines the role of generative AI, particularly ChatGPT, in enhancing education through systems thinking (ST) across various contexts. It assesses ChatGPT's performance in terms of accuracy, helpfulness, and reliability in responding to educational inquiries, revealing that the tool significantly aids in developing ST skills by offering immediate feedback, facilitating self-paced learning, and fostering interactive engagement among students. Despite these benefits, the document acknowledges several challenges, including instances of inaccuracies in ChatGPT's responses, concerns regarding data security, and the potential risk of students becoming overly dependent on AI-generated information. This necessitates a critical approach to evaluating the outputs produced by generative AI tools in educational settings. Overall, the findings suggest that while generative AI has the potential to positively impact learning and teaching processes, careful consideration and integration into educational practices are essential to mitigate associated risks.

Key Applications

ChatGPT

Context: K-12 and higher education students learning about systems thinking.

Implementation: ChatGPT was utilized to answer prompts related to systems thinking, causal loop diagrams, and system dynamics.

Outcomes: ChatGPT provided largely correct and helpful responses, facilitating personalized learning experiences and enhancing students' systems thinking skills.

Challenges: Inaccuracies in responses, risk of cheating, variability in answer quality, and potential over-reliance on AI.

Implementation Barriers

Technical Limitations

ChatGPT's inability to process images or logical diagrams and variable response quality.

Proposed Solutions: Use plugins like 'Show Me' and 'Code Interpreter' for visualization and model creation.

Data Security

Concerns over user prompts being used for further model training.

Proposed Solutions: Implement strict data privacy measures and educate users about data security.

Reliability and Trust

Inconsistencies in ChatGPT's responses can lead to misinformation and misconceptions among students.

Proposed Solutions: Encourage critical thinking and verification of information, and provide guidance on prompt formulation.

Over-reliance on AI

Potential for students to depend too heavily on AI, diminishing the role of teachers.

Proposed Solutions: Maintain a balanced approach where AI supplements rather than replaces teacher-led instruction.

Project Team

Holger Arndt

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Holger Arndt

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies