Results from the Undergraduate Perspectives on AI at Â鶹Ãâ·Ñ°æÏÂÔØBoulder Survey

Survey

Photo by on

In Spring 2025, theÌýEducational Technology Research Assistants (ETRA) conducted a mixed-methods research study to gain insight into the perspectives of Â鶹Ãâ·Ñ°æÏÂÔØundergraduate students on generative AI. The researchers asked the following questions:

  • How and why are Â鶹Ãâ·Ñ°æÏÂÔØBoulder undergraduates actually using AI?
  • What do they think about this technology?
  • And what policies and practices do students want their instructors and the broader institution to implement going forward?

Based on a synthesis of 2,992 survey responses (post data cleaning) and 17 interviews, the general findings and implications are as follows:Ìý

ÌýÌýFindings
  1. Diverse Student Perspectives
    • Undergraduate students hold a wide variety of perspectives on generative AI.ÌýÌý
  2. Widespread, Responsible Use
    • While generative AI use is widespread amongst undergraduate students, it’s generally not used for malicious purposes or to cut corners. Most respondents reported using it for helpful or supportive tasks, such as brainstorming, studying, or helping when they get stuck.
  3. Limited Academic Use
    • 68% of respondents use AI for less than a quarter of their academic work.
  4. Support During Difficult Moments
    • Students report using AI in their academic work when they are stuck on an assignment or need help understanding material, or when in-person help is not available.
  5. Reasons for Not Using AI
    • Not all students use AI. Reasons cited for abstaining are: concern about the accuracy of the output, wanting to respect their faculty member’s policies, and concerns that it will negatively impact their learning.
  6. Desire for Faculty Guidance
    • Students want to learn about AI from Â鶹Ãâ·Ñ°æÏÂÔØBoulder educators.
  7. False Accusations Happen
    • 10% of respondents have been falsely accused of submitting AI-generated work.
  8. Call for Clear Communication
    • Respondents reported that they wanted clear, open, and honest communication. This includes disclosure about how educators are using AI and guidance on students’ acceptable uses of AI.
  9. Skepticism About Grading with AI
    • While the majority of respondents are okay with their faculty member’s use of AI detectors, they are not comfortable with using AI to grade their assignments.
  10. Preference for Choice
    • Like faculty members, students want the ability to choose when they use AI in their classes.Ìý
ÌýÌýImplications
  1. Honor Student Perspectives
    • Respect students’ varying attitudes toward generative AI.
  2. Increase Learning Support
    1. Consider providing more learner support resources for students, to deter them from using generative AI in ways that might harm their learning.
  3. Avoid Overreliance on AI Detectors
    • Current findings on the accuracy limitations of AI detectors suggest that they should not be used to identify academic integrity violationsÌý(SeeÌý,Ìý;Ìý)
  4. Take a Learning-Centered Approach
    • When misuse of AI is a concern, consider approaching students in a non-confrontational manner andÌý
  5. Build Policies with Students
    • Foster transparency in the classroom byÌý with students. For first-year students, this may be the first time they’ve encountered this kind of exercise, so expect them to be reticent or think the exercise is ‘weird.’Ìý Having students participate in the policy-making process provides them with agency, responsibility, and promotes ownership of their actions.
  6. Model Disclosure
    • and ask students to do the same.
  7. Avoid AI-Based Grading
    • Consider refraining from using AI to grade student work.

Detailed results will be published on a rolling basis throughout the Fall 2025 semester. Ultimately, our goal is to inform future initiatives and interventions related to generative AI on the Â鶹Ãâ·Ñ°æÏÂÔØ campus. If you’d like to request a presentation or discuss this research, please contactÌýJacie Moriyama.Ìý ÌýÌý

Past Presentations