The Student GenAI Paradox

Human-centered research on GenAI usage and perspective from grad students at Information School from the Pratt Institute

A group of 12 graduate students collaborated with Dr. Irene Lopatovska at the Pratt Institute to explore the attitudes, beliefs, and behaviors of graduate students using generative AI for academic work. The study aimed to examine the relationship between students’ emotions and their experiences with AI.

GOAL

Explore usage, emotion, ethics, and expectations from GenAI use

Research

How do students use GenAI tools, and how do they feel about them?

Primary method

Survey to understand student's expectations with GenAI

Secondary Method

Focus Groups to better understand students' mindsets and personal experience with GenAI

Focus Group DaTA
Drawing of One Participant's Mental Model of How GenAI Works

In the focus groups, we asked participants to draw their mental models of Gen AI.

These provided a visual insight to the participant's mental understanding of GenAI. Many expressed negative views: environmental impact, art-theft, or ones like above that equate AI to shit. Others focused on the positives: chatGPT's translation abilities, the relationship between the user and the interface, and the ability to enhance one's work efficiently.

Results

80% of students reported using GenAI

Most common tasks: 
brainstorming (67%) and 
writing (62%)

30% feel GenAI increases school engagement while 61% say it doesn’t

37% felt GenAI supports creativity

93% of students surveyed are concerned about AI bias

10% would consider trusting GenAI without verification

59% expect their usage of GenAI to grow

GenAI use feels inevitable, even for skeptics

Only 3% think it’s always ethically fine; 60% say “it depends”

51% of respondents were outright uncomfortable using GenAI for schoolwork

47% are cautious to admit using GenAI

Some feel shame or worry about authorship

Our research formed 3 core themes

Pragmatic Use vs. Emotional Distance

Prompt engineering is a skill—it’s important to learn it now

Many students felt ashamed of using AI but justified it due to pressure to complete assignments perfectly and on time. They wanted more time for writing and brainstorming without pressure. While students acknowledged ethical concerns, they distanced emotionally from AI, using it as a tool in their academic work. This theme highlights the personal nature of each student’s relationship with AI, emphasizing the need for ongoing discussions about its impact on academic environments.

Ownership, Shame, and Ethics

Using GenAI...it’s still our idea, [GenAI is] just enhancing it

I don’t use GenAI for brainstorming because...I want to own my idea

In a school of design, ownership of work is closely linked to an individual’s innovation. The use of AI presents conflicting views on the integrity of one’s work. Some students believe AI dulls their creativity, feeling it diminishes their personal touch, while others see AI as a valuable tool that enhances and supplements their work.

Ownership, Shame, and Ethics

Using GenAI...it’s still our idea, [GenAI is] just enhancing it

In a school of design, ownership of work is closely linked to an individual’s innovation. The use of AI presents conflicting views on the integrity of one’s work. Some students believe AI dulls their creativity, feeling it diminishes their personal touch, while others see AI as a valuable tool that enhances and supplements their work.

Inspired by these recurring themes, I IDEATED

Three Hypothetical AI Tools Designed for Students

Instant, Ethical Writing Support

51% students felt uncomfortable using ChatGPT for schoolwork. Still, they used it to keep up with deadlines and academic pressure. A school-approved AI writing tool could guide students to improve their writing without giving them the answers, helping them learn while using AI responsibly.

After-Hours AI Librarian Support

In focus groups, students shared that while they use tools like ChatGPT for writing and research, many still prefer the guidance of a librarian. However, limited library hours—especially for those working late at night—can be a barrier. An AI-powered librarian available after hours could provide students with research support when they need it most, while still encouraging independent inquiry and the use of credible sources.

Contextual Translation and Academic Thesaurus Tool

In the focus groups, numerous students reported using AI tools like ChatGPT to get better translations and definitions that standard tools like Google Translate or dictionaries often miss. By describing the type of word or tone they need, students get more accurate and academically appropriate suggestions from the AI.

Takeaways

Across the board, students expressed a clear need for stronger guidance—from professors and administrators alike—on how AI should be used to support academic work.

This human research study revealed the conflicting attitudes students hold toward generative AI in their coursework. Many recognized its environmental impact and felt it reduced their own creative input, yet still relied on it to cope with academic stress and deadlines. Others saw GenAI as an inevitable part of the future and used it freely.

Next Steps
To deepen our understanding of GenAI’s role in academia, the study is now expanding to include faculty perspectives, especially as they confront the challenge of evaluating work that may include AI-generated content.

What I Learned
Working with a team of engaged student researchers on such a timely, complex topic sharpened my research skills and broadened my understanding of how emerging technologies shape the academic experience. It also gave me valuable insight into the diverse ways students navigate—and negotiate—the presence of AI in their education.