article · Disciplinary and Interdisciplinary Science Education Research
Integrating ChatGPT into higher education STEM teaching can improve conceptual comprehension and correct common scientific misunderstandings. A quasi-experimental trial involving 120 first-year engineering students evaluated the use of ChatGPT alongside a Constructivist Inquiry-Based Learning Prompting framework during an introductory thermodynamics course. Students using the generative AI tool demonstrated significantly stronger gains in conceptual understanding compared to peers who received traditional instruction. The intervention reduced qualitative misconceptions concerning complex topics such as entropy and internal energy. Nonetheless, quantitative errors persisted, highlighting constraints in handling advanced reasoning and numerical calculations. Weekly surveys revealed high student satisfaction with the usability and instructional value of the tool. Importantly, optimal learning gains were associated with targeted rather than continuous reliance on the software, demonstrating the practical value of guided generative AI integration in educational environments, particularly within resource-constrained contexts.
Teaching complex engineering subjects often struggles with abstract concepts and student misunderstandings. This research demonstrates how guided generative AI frameworks can act as effective supplementary teaching aids, particularly in resource-constrained settings. By highlighting both the benefits for conceptual grasp and the limitations in numerical problem-solving, it provides educators with realistic guidance on balancing AI tools within science and engineering curricula.
The tested prompting framework has direct application for higher education institutions, educational technology providers, and STEM curriculum developers seeking to incorporate generative AI into learning management systems. As an applied and tested teaching method evaluated in a classroom setting, the approach is ready for adaptation into university coursework. Further commercial or operational deployment will require safeguards or complementary tutoring modules to address the identified limitations in numerical calculation and advanced mathematical reasoning.
AI-generated from the published abstract. Always read the original work before citing.
Abstract Advancements in artificial intelligence (AI), particularly generative AI models such as ChatGPT, offer transformative opportunities to enhance educational practices in STEM disciplines. Thermodynamics, a fundamental subject in engineering education, presents significant challenges due to its abstract nature and common misconceptions. This study investigates the effectiveness of integrating ChatGPT as a supplemental pedagogical tool, guided by a constructivist inquiry-based approach using the Constructivist Inquiry-Based Learning Prompting (CILP) framework, to enhance conceptual understanding and address misconceptions in an introductory thermodynamics course for first-year Moroccan engineering students. A quasi-experimental design was used, with 120 students equally divided into control and experimental groups. The control group received traditional instruction, whereas the experimental group received ChatGPT-assisted instruction. Conceptual understanding was measured using pre- and post-tests, while student perceptions and acceptance were collected via weekly surveys. Results showed that the experimental group significantly outperformed the control group, exhibiting greater improvements in conceptual understanding and a reduction in qualitative misconceptions, particularly related to entropy and internal energy. However, some quantitative misconceptions persisted, underscoring ChatGPT’s limitations in advanced reasoning tasks, problem-solving, and numerical calculations. Students reported high satisfaction with ChatGPT’s usability and instructional support. Moreover, targeted use of ChatGPT, rather than frequent reliance, correlated with optimal learning outcomes. These findings underscore ChatGPT’s potential to enhance STEM education within inquiry-based, constructivist learning environments and provide evidence for the effective integration of generative AI tools to improve learning outcomes, particularly in resource-constrained settings.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1186/s43031-025-00125-z
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
