Integrating Generative AI into Environmental Engineering Sustainability Education: The Lorax as a Classroom Case Study

Introduction

Engineering education has evolved over time, much as the tools for the trade and implementation of engineering have evolved. What was once state-of-the-art is now obsolete, for example, the transition from the slide rule to the electronic calculator. With the advent of these new tools, educational methods have changed and adapted. Computer programming is now considered an essential part of an engineering curriculum, when a hundred years ago that would have been an unthinkable concept. These disruptive technological shifts changed and, in many instances, have improved how engineering is done. At the same time, new technologies are often approached with fear—the introduction of elevators as an example—and sometimes that fear is justifiable, often relying on the improvements of the technologies—such as elevator brakes—before wide scale acceptance. Technologies themselves are generally never good nor bad inherently; however, how they are employed by humans greatly influences their narrative.

Generative artificial intelligence (AI) tools such as ChatGPT represent a new evolution in the constant turning of new technologies, and considering their role in engineering education. There has been much written in the last few years around these technologies, both highlighting the possible benefits and challenges (Alasadi and Baiz, 2023; Qadir, 2023; Yusuf et al., 2024). Regardless of the different stances, it is important to see these AI tools as they are, as tools which may both have the potential to be used for the benefit and destruction of humankind. The same could be said of dynamite, it has both the potential to enable the usage of new critical resources through applications such as mining, and the potential to kill people (Sachs, 2024). The question now is how should tools such as generative AI be applied in the context of environmental engineering and sustainability education and beyond in order to facilitate learning? The literature around AI usage in an educational context is constantly evolving (AAC&U and Elon University, 2026; Baek et al., 2024; Flaherty, 2025; Stripling, 2025).

In the application of sustainability and engineering education, the challenge is how to use these tools to enable deeper thinking of sustainability topics instead of using them to replace critical thinking. This work explores the student critique of a ChatGPT response to a class-relevant prompt centered around a reading of and an essay based on Dr. Suess’s The Lorax. The Lorax, a children’s book first published in 1971, explores many sustainability topics through a story based around resource use and deforestation.

The overarching goal of this work is to generate insight as to how students perceive AI and what value they see of AI in a sustainability educational context. In particular, this work also provides a time-based snapshot (circa 2023) of these student perceptions, which are likely to evolve over time with the proliferation of AI throughout the educational landscape. We know that students perceive some value to AI in the educational context, as many students are using AI either in a sanctioned and structured manner in their coursework or simply as a shortcut (Stripling, 2025; Terry, 2024).

Study Design

A prompt regarding reflection on The Lorax was entered into ChatGPT on August 25, 2023: “The Lorax is a well-known children’s book by Dr. Seuss, with the text listed below. The assignment is to read the text critically, drawing from what you have learned in this course. Then, write a 1,500-word analysis of the larger point that Dr. Seuss is attempting to make. Feel free to draw on examples from current events and other literature.” ChatGPT provided a response, which is provided in full in the Supplemental Data. Students were tasked with revising the AI-generated response to fit the parameters of the course assignment and answering five guided reflection questions based on their previous and current experience with generative AI (including this assignment):

Q1: What did you like about the AI-generated response?

Although this is a regular course assignment, it also constitutes human subjects research, and thus the protocol of students choosing whether or not to have their assignment included in the research dataset was governed under the University of Wisconsin-Madison Institutional Review Board (IRB) on Human Subjects Research: 2023:1336. Of the 38 students in the course 33 number opted in to the research. Details of the text analysis and data processing methods may be found in the SI.

Results

First, we focused on positive and negative sentiment words (Fig. 1a and b) with respect to the AI-generated essay. With respect to Q 2 , negative sentiment towards were revealed towards the AI-generated prompt. We found that students regarded at times the AI-generated response as superficial, lacking critical analysis, and did not align with the supporting themes (i.e., those relating to The Lorax, for instance, the themes relating to the tragedy of the commons). Additionally, Figure 1c and d revealed a major discussion from students around data inclusion. Which went back to the negative sentiment of why students did not like the ChatGPT response, as it did not include data that they hoped to find, as compared to in-class discussions they had participated in.

OPEN IN VIEWER

FIG. 1.

(a) positive sentiment analysis words, (b) negative sentiment analysis words, (c) Q2 topic 1, (d) Q2 topic 2, (e) Q3 notable topics, (f) Q4 topic 1, (g) G4 topic 2, (h) Q5 topic 1, (i) Q5 topic 2.

Note that, while Q 1 focused on things students liked about the AI-generated response, we did not find very notable topics of discussion, as most revolve around The Lorax discussion rather than general opinions about the AI as a tool (Fig. 1a).

In Q 3 —that asked the best ways to employ AI for essay prompts—we saw that students predominantly focused on a single topic shown in Figure 1e Students was that AI was most useful in generating ideas and themes, help in brainstorming, checking papers, and responding to essays.

In Q 4 —where students reflected on the role of AI in education—we found two notable discussion topics, seen in Figure 1f and g. Students reflected that one notable role of AI in education was to personalize the learning experience, save time, and provide critical thinking. This was of specific interest to educators at open avenues to personalize the learning material they offer through AI tools. Further, students found that AI’s role could be useful in providing information, writing code, and learning material. Interesting, students as well reflected on AI’s ability to complete assignments and its use in that sphere. While this was expected, it reflects on the dilemma educators are exposed to when acknowledging the role of such tools in helping students and perhaps trivializing learning expectations (i.e., assignments, exams, etc.). And highlights the potential for students using AI in place of developing their critical thinking skills.

Finally, in Q 5 —which asked about previous experience with AI—we saw two notable topics as well, Figure 1h and i. Results here were rather expected. Students mostly had experience with ChatGPT, and they have used such AI tools to write assignments, research topics, and solve various problems. Given that ChatGPT was first released in November 2022 and became a fast-growing technology product that is an anticipated response.

Discussion

The prevalence of generative AI tools, present a transformative opportunity towards its integration into engineering education. This work studied student reflections on multiple questions regarding AI usage and critique of ChatGPT-generated response on Dr. Suess’s The Lorax.

Our findings showed a general positive sentiment among students towards such AI tools. Such sentiment was mostly influenced by students’ appreciation of AI’s capability to provide personalized learning, save time on tasks, generate ideas, and help in writing or coding. Yet, students as well expressed a negative sentiment driven by superficial analysis, confusion, and problems with the AI output. For faculty teaching similar classes, this work provides one example of how to have students engage critically with AI-generated essays, and at the same time create a snapshot of student perceptions. This same type of exercise could be done utilizing different readings or even film offerings. At the same time, a challenge now that AI models have evolved, is ensuring that students are editing the AI response, and not simply having AI revise the initial prompt.

AI is a tool, and like all tools, its potential to change the status quo will depend on how it is deployed. At the same time, AI has the potential to be a very disruptive technology, not only in educational sphere, but also beyond. Regardless of how AI is remembered in the history books for changing education, it is and will change education. This work, conducted in 2023, provides a snapshot in time of student perceptions of the utility of AI in the context of sustainability and engineering education. Since 2023 AI tools have improved exponentially not only in their performance but also availability. More recently there are many accounts of how AI tools are being utilized during college coursework that necessitates a rethinking of traditional assessment activities (Truex, 2026; Gutkin, 2026), as these models have evolved.

Conclusions

AI represents the next evolution in technology, with the potential to be disruptive in an educational context. We have seen disruptions before due to evolutions from the slide rule to handheld calculators. This work presents one classroom case study involving The Lorax and AI. Future snapshots in time should continue to be conducted in order to understand how the evolution of AI is influencing student perceptions of its usefulness in a sustainability engineering context.

Data Availability

The data generated and/or analyzed during the current study are not publicly available for legal/ethical reasons but are available (as aggregate data) from the corresponding author upon reasonable request.

Code Availability

The code used in the analysis was developed in Matlab R2024a and is provided.

Supplemental Material