Written By: Jesan Ahammed Ovi (My PhD Student in Computer Science at the Colorado School of Mines)
As Generative AI tools like ChatGPT, DALL·E, and GitHub Copilot rapidly enter classrooms and labs, one question looms large: How are students actually using them—and how do they feel about it?
Our study at the Colorado School of Mines (a small, engineering-focused R1 university) tracks student adoption of GenAI from Spring 2023 to Fall 2024. The data reveals a complex story—a mix of enthusiasm, concern, and existential unease.
We surveyed the entire population of engineering students, to explore their usage, motivations, ethical concerns, and even their thoughts on AI-induced doom (yes, really). Here’s what we found.
GenAI Use Is Up—and So Is Concern
First, the obvious: student adoption is increasing. We observed that regular users (who used LLMs weekly) and superusers (who used LLMs daily or more often), have significantly increased from 2023 by 9.8% and 3.5% respectively. LLM-powered chatbots are now a regular part of many students’ academic routines. But students aren’t blindly optimistic. Alongside excitement, we found deep and nuanced concerns.
Figure 1: Students’ top concerns regarding GenAI in 2024 (n = 862). X-axis: frequencies normalized by # of respondents per dept. Cluster.
Students worry that AI tools might:
- Spread misinformation
- Be trained on biased or unreliable data
- Encourage plagiarism or over-dependence
- Job displacement
This tension—between opportunity and risk—is shaping how students want to engage with GenAI tools.
Motivation: Learning First, Curiosity Second
On the positive side, students primarily use GenAI tools to learn better and improve their work quality. The top motivations include:
- Gaining deeper understanding of course material
- Enhancing quality of the work
- Exploring new tech out of curiosity
This suggests that GenAI is not simply a shortcut—but often a supplemental learning tool.
What’s Your p(doom)? Students Reflect on AI Catastrophe
Among our most striking findings was students’ response to the question: “What is your estimate for p(doom)”—the probability that AI will cause catastrophic harm to humanity?
Yes, this is a real concept used in AI safety debates—and students engaged with it seriously.
Figure 2: Students’ estimation of P(doom) – how likely AI will cause catastrophic harm. X-axis: P(doom) estimates from 0 (no chance) to 100 (certain catastrophe).
Students are divided into two groups:
📉 One group, relatively smaller but significant, estimates p(doom) at around 20%. That means 1 in 5 odds that AI could end humanity—already an alarming signal.
📈 The second and larger group centers their estimates near 60%, suggesting a dominant fear that catastrophic AI harm is more likely than not.
This isn’t just academic pessimism. It reflects something deeper: Students aren’t only adopting GenAI tools for their coursework—they’re grappling with AI’s future impact on society, science, and survival.
What Should Institutions Do?
Based on our findings, institutions must respond not just to how students use GenAI—but also to how they feel about its future. We recommend that educators and administrators:
- Establish clear and nuanced policies for GenAI use, particularly around ethics, plagiarism, data privacy, and appropriate academic use.
- Equip both students and faculty with training and resources to better understand GenAI’s capabilities, limitations, and best practices.
- Go beyond the hype. Facilitate critical conversations about GenAI’s broader societal risks—not just its rewards.
- These conversations should include topics like:
🌍 Digital divide and access inequality
♻️ Environmental and climate impact of large-scale AI
🏛️ Concentration of power among a few AI tech giants
⚖️ Bias and fairness in AI training data and outcomes
By fostering these discussions and designing support structures, institutions can empower the next generation of engineers and scientists to become not just users, but thoughtful stewards of AI technologies.
A Call to the Community
This study focuses on one engineering school—but these questions apply everywhere:
- What do students in other universities think?
- What’s happening in disciplines beyond engineering?
- How do faculty perspectives align—or clash—with student views?
Read More or Get Involved
Want to learn more? Curious about collaborating or replicating this study at your institution?
📄 Read the full paper
📬 Reach out—we’d love to compare notes! Email Jesan Ahammed Ovi at jesanahammed_ovi@mines.edu or Prof. C. Estelle Smith at estellesmith@mines.edu.
C. Estelle Smith, Ph.D. 