💻 computer science

To Use or Not to Use: Investigating Student Perceptions of Faculty Generative AI Usage in Higher Education

This study investigates student perceptions of faculty Generative AI usage in higher education, revealing that while a significant portion of students support its dual adoption, many remain skeptical about the technology's reliability and fear that faculty overreliance could undermine critical thinking and pedagogical quality.

Original authors: Jie Gao, Jiayi Zhang, Dan Chen

Published 2026-03-30

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Jie Gao, Jiayi Zhang, Dan Chen

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the classroom as a busy kitchen. For a long time, the rule was simple: Students are the ones cooking their own meals (doing assignments), and Teachers are the chefs who taste the food, give feedback, and decide if it's ready to serve (grading).

Now, a new, super-fast robot chef called Generative AI (GenAI) has entered the kitchen. It can chop vegetables, mix sauces, and even write recipes in seconds.

This paper is like a survey asking the students: "How do you feel about this robot?" But here's the twist: The researchers didn't just ask, "Do you want to use the robot to cook your own dinner?" They also asked, "Do you want the teacher to use the robot to grade your dinner?"

Here is the breakdown of what they found, using simple analogies:

1. The Great "Double Standard" (The Asymmetry)

The study found that students have a double standard, kind of like a parent who says, "You can use a calculator for your homework, but I'm not letting you use one to grade my taxes."

The "I'll Do It" Crowd: About 56% of students said, "Yes, I should be allowed to use the AI robot to help me cook my assignments." They see it as a helpful tool to save time.
The "You Can't Do It" Crowd: However, only 37% of students said, "Yes, the teacher should use the AI robot to grade my work."

The Result: There is a big gap. Many students are okay with using the robot for themselves, but they get nervous when they think the teacher might let the robot do the grading or teaching.

2. The Four Types of Students

The researchers sorted the students into four "flavors" based on their opinions:

The GenAI Optimists (31%): These are the tech-lovers. They say, "Let the robot help me cook, and let the robot help the teacher grade! Everyone wins!"
The Non-Supporters (37%): These are the traditionalists. They say, "No robots allowed. I want to cook my own meal, and I want the human teacher to taste it."
The Student Support Group (26%): These are the "Me first" group. They say, "I want to use the robot to help me, but the teacher must do the grading themselves." (This is the biggest group showing the double standard).
The Faculty Support Group (Small %): A rare group who say, "I don't want to use the robot, but I think the teacher should use it to grade."

3. Why Are Students Worried? (The 4 Big Fears)

When the students who didn't want the teacher to use AI were asked why, their answers fell into four main buckets. Think of these as the "Four Horsemen of the Classroom Apocalypse":

The "Lazy Chef" Fear (Professional Concern):
Students worry that if teachers use AI, they will stop trying. It's like if a chef stopped tasting the food and just let the robot decide if it's salty enough. Students fear teachers will become "lazy," lose their passion, and stop connecting with them personally.
The "Fake News" Fear (Quality Concern):
Students know the robot sometimes lies (hallucinates). They are terrified that the teacher will use the robot to grade their work, and the robot might give a "C" to a masterpiece or a "B" to a disaster because the robot got confused. They don't trust the robot's brain.
The "Brain Atrophy" Fear (Pedagogical Concern):
This was the biggest worry (37% of concerns). Students fear that if teachers rely too much on AI, the teachers themselves will stop thinking critically. It's like a muscle that gets weak if you don't use it. If the teacher stops thinking deeply because the AI does it for them, the students won't learn deep, critical thinking skills either. They fear the "human spark" of teaching will die out.
The "Unfair Judge" Fear (Ethical Concern):
Students worry about cheating, bias, and privacy. They are worried the robot might be biased against certain people or that it will accidentally steal ideas (plagiarism).

The Bottom Line

The main takeaway is that students aren't just "for" or "against" AI. They are nuanced.

They are okay with AI being a tool (like a calculator or a spell-checker) to help them learn. But they are very skeptical of AI becoming the judge (the teacher). They want to know that a real human is still in the kitchen, tasting the food, and making sure the lesson is real, fair, and full of human connection.

In short: Students want the robot to help them cook, but they want a human to be the one who decides if the meal is good.

1. Problem Statement

While Generative AI (GenAI) has rapidly integrated into higher education, existing research predominantly focuses on regulating student usage. There is a significant gap in understanding how students perceive faculty adoption of GenAI for instructional design, teaching support, and assessment (grading).

The core problem addressed is the potential asymmetry in student attitudes: students may view GenAI as a legitimate learning aid for themselves while simultaneously resisting its use by instructors for teaching and grading. This asymmetry poses risks to instructional design, assessment transparency, and institutional policy if not understood. The study seeks to quantify this divergence and identify the specific pedagogical, ethical, and professional concerns driving student resistance to faculty GenAI usage.

2. Methodology

The study employs a mixed-methods approach utilizing a secondary analysis of an open-access dataset hosted on Harvard Dataverse.

Data Source & Participants:
- Sample Size: $N = 156$ undergraduate and graduate students.
- Demographics: Students from universities in the US, Canada, UK, Australia, and France, spanning diverse majors (Arts, CS, Engineering, Medicine, Law, etc.).
- Data Collection: Anonymous online survey via Google Forms.
Key Variables & Questions:
- Quantitative: Four Likert-style items measuring attitudes toward:
  1. Student use of GenAI in assignments.
  2. Faculty use of GenAI in teaching.
  3. Faculty use of GenAI for grading.
  4. Open-ended concerns regarding faculty GenAI use.
Analytical Techniques:
1. McNemar's Test: Used to determine if there is a statistically significant difference (asymmetry) between students' acceptance of their own GenAI use versus faculty GenAI use for grading.
2. Cluster Analysis (Categorization): Students were classified into four distinct groups based on their binary responses (Support/Oppose) to student vs. faculty usage:
  - GenAI Optimists: Support both.
  - Student Support Group: Support student use, oppose faculty use.
  - Faculty Support Group: Oppose student use, support faculty use.
  - Non-supporters: Oppose both.
3. Thematic Analysis: Applied to 130 open-ended responses regarding concerns. A coding scheme was developed based on prior literature, resulting in four non-mutually exclusive themes: Professional, GenAI Quality, Pedagogical, and Ethical. Inter-coder reliability was $\approx 92\%$ .

3. Key Contributions

Shift in Focus: Moves the research lens from student regulation to student perceptions of faculty adoption, addressing a critical blind spot in current GenAI literature.
Empirical Evidence of Asymmetry: Provides statistical proof that student acceptance of GenAI is context-dependent, revealing a significant divergence between personal utility and trust in institutional application.
Taxonomy of Student Attitudes: Proposes a four-group classification model (Optimists, Student Support, Faculty Support, Non-supporters) that captures the heterogeneity of student views, rejecting a simple pro/anti dichotomy.
Qualitative Insight into Resistance: Identifies that student resistance to faculty usage is driven less by technical limitations and more by pedagogical and professional concerns (e.g., erosion of critical thinking, faculty laziness).

4. Key Results

A. Asymmetry in Acceptance (RQ1)

Statistical Significance: A McNemar's test revealed a statistically significant divergence in attitudes ( $\chi^2 = 16.820, p < .001$ ).
Usage Disparity:
- 56.4% of students support their own use of GenAI for assignments.
- Only 37.2% support faculty use of GenAI for grading.
Group Distribution:
- Non-supporters (37.2%): Oppose GenAI use in both contexts.
- GenAI Optimists (30.8%): Support GenAI use in both contexts.
- Student Support Group (25.6%): Support student use but oppose faculty use (demonstrating the asymmetry).
- Faculty Support Group (Minority): Oppose student use but support faculty use.

B. Primary Concerns Regarding Faculty Use (RQ2)

Thematic analysis of open-ended responses revealed four dominant concern categories:

Pedagogical Concerns (Most Prominent - 37%):
- Students fear faculty overreliance on GenAI will create a "futile cycle," reducing faculty critical thinking skills.
- Concerns that AI will distance the student-faculty relationship, reduce teaching individuality, and inhibit student creativity/originality.
GenAI Quality Concerns (79% of those citing quality issues):
- High skepticism regarding the validity, reliability, and accuracy of GenAI outputs.
- Specific fears of "hallucinations," inaccurate grading, and the propagation of false information.
Professional Concerns (21%):
- Perception that using GenAI for teaching/grading constitutes professional dereliction or "laziness."
- Fear that faculty will abdicate instructional responsibility and lose genuine passion for teaching.
Ethical Concerns:
- Issues regarding academic integrity (plagiarism), algorithmic bias, and data privacy.

5. Significance and Implications

Policy Formulation: Institutions cannot assume that because students use GenAI, they will accept it from instructors. Policies must address the specific "trust gap" regarding faculty usage.
Pedagogical Design: The findings suggest that for GenAI to be accepted in teaching, it must be framed as a tool that augments rather than replaces human interaction and critical thinking.
Faculty Development: Instructors need training not just on how to use GenAI, but on how to communicate its role to students to mitigate fears of "laziness" or reduced educational quality.
Future Research: Highlights the need to investigate the underlying reasons for the asymmetry (institutional, social, or individual factors) and how these perceptions evolve as GenAI technology matures.

Conclusion: The study concludes that students hold nuanced, role-sensitive views on GenAI. While a significant portion (31%) is optimistic, a large segment (37%) is resistant to GenAI in any form, and another quarter (25.6%) specifically rejects faculty usage. Successful integration requires addressing the pedagogical integrity and human element of education rather than focusing solely on technical efficiency.