(376h) Artificial Intelligence (AI) Assisted Thermodynamics Learning

In this moment of proliferation of artificial intelligence (AI), major questions exist as to how to best approach these advancing technologies from a pedagogical standpoint. Ignoring them entirely has the potential to leave students at a disadvantage for their future work and careers, however, unrestricted use has the potential for issues of academic integrity. This study takes the approach that there can be a middle ground in which the student drives the AI in order to allow them to think more about the overall problem at hand, rather than the granularities of writing complex code. This is particularly evident in the area of thermodynamics.

Within thermodynamics, a limiting factor in solving certain complex problems is the student’s ability to carry out tedious mathematical calculations in addition to their engineering analysis. This study utilizes ChatGPT to assist students in writing MATLAB code without needing an independent and extensive command of MATLAB coding language. ChatGPT working in collaboration with the students can produce the general structure of the required code for the student to then customize based on their engineering analysis and specific thermodynamics question. This allows us to imagine a future where the focus is on problem set up, rationale, critical thinking, and engineering analysis, without being limited by tedious calculations. This teaching methodology was rolled out in a graduate level thermodynamics course utilizing ChatGPT to write MATLAB code for systems of ordinary differential equations, molar volume and fugacity calculations of pure gases and binary gas mixtures, vapor liquid equilibrium calculations, etc. Students were surveyed both before and after taking thermodynamics with AI to assist in coding.