University of Hawaiʻi at Hilo

AI Perceptions & Attitudes Survey

Comprehensive analysis of how the UH community perceives artificial intelligence, its impact on education, and future implications.

🎯 Core Objective Longitudinal study of the "AI Ecosystem" on campus—measuring awareness, acceptance, and usage
👥 Principal Investigators Sukhwa Hong, Ph.D. (UH Hilo) · Mahdi Belcaid, Ph.D. (UH Mānoa) · Jason Leigh, Ph.D. (UH Mānoa)
🔬 Methodology Explanatory Sequential Mixed-Methods · IRB #2024-00045
🤝 Partnership San Diego State University
💡 Key Finding "The Usage Paradox" — Student AI usage declined twice as much as faculty usage (54→39 vs 54→47, p=.004), revealing a "Fragmented Reality" with divergent responses to AI adoption.
--
-- Responses
-- % of Total
849
Total Responses
10
UH Campuses
22%
Paid AI Subscribers
2
Survey Years
88/100
Verify AI Accuracy (normalized)

📊 Key Research Findings

🔄 Interaction Effect p=.004

The Usage Paradox

Students Declined More Than Faculty
Faculty Usage 54 → 47 ↓
Student Usage 54 → 39 ↓↓
📉 p < .001, d=0.34

Hype Cycle Crested

From Inflated Expectations to Pragmatic Realism
Future Expectations 54 → 39 ↓
Perceptions & Attitudes 48 → 35 ↓
📊 2025 Cross-Sectional

The 2025 Reality Gap

Faculty Scored Higher on ALL 5 Constructs
Biggest Gap: Awareness 58 vs 39 (+19)
Skills Gap 54 vs 45 (+9)
💰 5× Growth

Economics of Access

Surge in Paid AI Subscriptions
Paid Subscriptions 4.3% → 21.9% ↑
Faculty vs Students (2025) 24% vs 19%
✅ Consensus > 85/100

Human-in-the-Loop

Universal Agreement on AI Verification
Faculty: Verify AI 90 / 100
Students: Verify AI 85 / 100
🔍 Consensus > 75/100

Ethical Transparency

Algorithms Should Be Transparent
Students 74 / 100
Faculty 77 / 100
🎯

Conclusion: "The Fragmented Reality"

We are not witnessing a uniform adoption curve, but two distinct cultural responses. Faculty show Cautious Integration (increasing usage, institutional support, normalizing professional application) while Students exhibit Disillusioned Withdrawal (decreasing usage, declining confidence, "shadow pedagogy").

🔍 Interactive Data Explorer

⚠️ Scale Note: 2024 used a 5-point scale (1-5) while 2025 used a 6-point scale (1-6). Scores are normalized to 0-100 for valid comparison.
2024
5-point Likert Scale
n = 351
-- Normalized
-- Raw Mean (1-5)
👎 Disagree (1-2) --%
Agree (4-5) 👍: --%
VS
2025
6-point Likert Scale
n = 498
-- Normalized
-- Raw Mean (1-6)
👎 Disagree (1-3) --%
Agree (4-6) 👍: --%
📊 Year-over-Year Comparison
Normalized Score Change --
Agreement Shift --
Combined Sample Size --

📊 Demographics Overview

Survey Participation by Year

Comparing 2024 and 2025 responses
Trend

Respondent Roles

Faculty/Staff vs Students
Distribution

UH Campus Affiliation

Responses across all UH campuses
Campus

Age Distribution

Respondent age brackets
Demographics

🤖 AI Attitudes & Perceptions

AI Sentiment Overview

Average agreement scores on 6-point Likert scale (1=Strongly Disagree, 6=Strongly Agree)
Key Metrics

Year-over-Year Comparison

Normalized scores (0-100) – 2024 used 5-pt scale, 2025 used 6-pt
Change

AI Subscription Status

Paid AI service usage
Usage

🎯 Key Gaps & Academics

The 2025 Reality Gap

Faculty vs Student scores across 5 constructs
Key Finding

Academic Disciplines

Student enrollment by field
Academic