Educators' Perceptions of Gen AI and its Influence on K-12 Education
Alexis Kane, Gena Lee, Ting Cheng, Yue Xu · HCDE 519, Autumn 2025
QUALATATIVE Research Study
K-12 Education
Generative AI
Executive Summary
How Are Educators Navigating the AI Era?
Research Questions
- RQ1: How do U.S. high school educators perceive Gen AI's influence on students' learning outcomes?
- RQ2: How are educators adapting teaching and curriculum in response to AI, particularly around student usage and AI literacy?
Research Methods
Two-Method Approach
Interviews
Four current U.S. high school teachers with 2–8 years of experience, in public and private schools across Seattle and San Francisco. Subjects taught included math, biology, chemistry, and computer science.
Netnography
Data collected from Reddit, Medium, YouTube, TikTok, and Instagram posts, comment threads, and video transcripts from educators. Each team member spent ~5 hours collecting data from assigned platforms, recorded in a shared spreadsheet.
Data Analysis Process
01
Data Preparation & Initial Coding
Semi-structured interviews were recorded and transcribed. Each team member then inductively coded interviews using ATLAS.ti, leveraging AI for initial pattern discovery.
02
Affinity Mapping
Codes were collaboratively affinity mapped, merging similar concepts and clustering them into broader themes.
An example of our affinity mapping process, grouping initial codes into emergent themes.
03
Codebook Development & Application
A comprehensive codebook, organized by theme families, was created and then applied deductively to our netnography data. This allowed for refinement as new codes surfaced.
Illustration of our final codebook structure, showing theme families and nested individual codes.
04
Analytic Memo Writing
Analytic memos documented reflections, emergent patterns, and potential inter-theme relationships, directly informing our final research findings and recommendations.
Finding 1 · RQ1: Impact of AI on Learning Outcomes
AI Impacts Students Non-Uniformly
Educators emphasized that AI's influence depends heavily on subject area, student skill level, learning needs, and access to resources.
Subject Area Divide
Humanities courses are seen as more vulnerable to AI misuse.
"It's not really good at solving or explaining math problems. Students…what they realize is that it's not really that helpful." (IP4)
Socioeconomic Disparities
Different levels of access to advanced technology, learning resources, and education on tech literacy create disparities in AI skill development.
“I think that it's supporting the students who maybe come from a family background that is a little bit better off.. Where from a young age, you're getting tutored, you are getting help, and you have a lot of the resources available. And so by the time you're in eighth or ninth grade, your literacy level is there, your critical thinking level is there. You have the foundation for it.”
Overall, AI is not a universally supportive tool, it amplifies existing inequities tied to subject matter, developmental stage, and socioeconomic status.
Finding 2 · RQ1: Impact of AI on Learning Outcomes
AI Threatens Critical Thinking & Literacy Development
Students over rely on AI rather than struggle through learning
“I think the presence of AI really takes away the time that students need to think for themselves… struggling with a problem, that’s part of learning.” IP3
Younger students are the most vulnerable
"More advanced or older students often self-correct after realizing AI negatively impacted their understanding of the material, something younger students potentially have not yet learned to do." IP4
Humanities teachers viewed as the most exposed because critical thinking in these courses is often built through writing and comprehension work
“You know, who cares about my opinion on the Great Gatsby? And the truth is, kind of nobody cares about your opinion ..but it's an exercise of critical thinking..” IP4
Finding 3 · RQ2: Educator Adaptations to Teaching & Curriculum
Educators Are Adapting with Two Broad Strategies
Reducing AI Cheating Opportunities
- Returning to pen-and-paper assessments
- Grading via discussion and presentation
- Restricting digital tools on school laptops
- Requiring revision history and source trails
- Minimizing homework; emphasizing in-class practice
Building Ethical AI Literacy
- Students critique and improve AI-generated work
- Debate exercises on AI outputs and limitations
- Transparent classroom AI use with guided reflection
- Teaching students to see where AI falls short
"Students will either learn to use AI incorrectly, or you can teach them to use AI to learn." - TikTok User 30
"You can't have AI literacy without actual literacy." - TikTok User 31
Finding 4 · RQ2: Educator Adaptations to Teaching & Curriculum
AI Exposes Deep Systemic Issues
District policies vague or nonexistent & teachers left to figure it out
Pressure to “teach AI” without time, training, or resources
Culture of grades over learning fuels AI misuse
Inequities widening between well-resourced and under-resourced schools
"As an educator, I think AI simply divulges how primitive and rigid our educational system is. It's one thing to pass down our knowledge. Whereas another to develop students into a self-evolving state, whom we failed miserably." Youtube9
“My administration talks a lot about ‘we should be thinking about AI’… but they don't say anything that's concrete… What do we need to teach them? What do we need to not teach them?” IP4
Core Finding
What Educators Are Asking For
U.S. educators lack the time, resources, and institutional support to properly address AI's influence. Most public school districts have yet to establish AI policies, leaving educators to navigate alone. AI amplifies these pre-existing resource gaps.
District-Level AI Policy
Clear, consistent guidelines from school districts and government, not left to individual educators to navigate alone.
Professional Development
Investing in structured professional development and training for educators to use AI ethically and responsibly in their classrooms.
Standardized AI Curriculum
Tech and AI literacy embedded into K-12 curriculum by districts , not an added burden on already overstretched teachers.
Limitations & Future Work
Expanding the Research Scope
Limitations
Narrow Participant Pool
All interview participants taught STEM subjects in Seattle and San Francisco in higher socioeconomic areas. This limits the generalizability of findings across subject areas and regions.
Recruitment Constraints
No participant compensation limited recruitment and capped interview sessions at one hour, restricting the depth of data collected.
Future Work
Diversify Participants
Expand recruitment across grade levels, subjects (especially liberal arts), and geographic regions to capture a wider range of educator perspectives.
Co-Design Workshops
With compensation, engage educators in longer sessions including co-design workshops focused on building AI and tech literacy curricula.
Empirical Cognitive Studies
Conduct empirical studies on the cognitive and developmental impacts of AI on student learning, current findings rely on educator observations and would benefit from direct measurement.
References & Appendix
Sources & Study Materials
References
Vieriu, A. M., & Petrea, G. (2025). The Impact of AI on Students' Academic Development. Education Sciences, 15(3), 343. doi.org/10.3390/educsci15030343
Zhai, C., Wibowo, S. & Li, L.D. (2024). Effects of over-reliance on AI dialogue systems on students' cognitive abilities. Smart Learn. Environ. 11, 28. doi.org/10.1186/s40561-024-00316-7
Lin, P., & Van Brummelen, J. (2021). Engaging Teachers to Co-Design Integrated AI Curriculum for K-12. CHI 2021. doi.org/10.1145/3411764.3445377