Rethinking How Students Interact With AI: Toward Human-Centered Learning
Authors: Yerin Seung; info@ciddl.org
Artificial intelligence (AI) has become a buzzword across education, with its potential for personalized learning, real-time feedback, and data-driven insights. However, as the conversation progresses, one crucial question remains: how are students actually interacting with AI in classrooms, and for whom are these systems truly designed? In our recently published article, Exploring Artificial Intelligence Integration and Student–AI Interaction in K-12 Education: A Scoping Review (published in the Journal of Computer Assisted Learning), we examined 70 empirical studies conducted over the past decade to map how AI is being integrated across school levels, subjects, and student populations. What we found tells a hopeful yet cautionary story about the future of AI in education.
A Rapidly Growing Field, But With Uneven Representation
The number of studies on AI in K–12 settings has grown dramatically in just a few years, mirroring the surge of interest since 2019. Most research is interdisciplinary, bridging education, computer science, and psychology, which is an encouraging sign that the field is maturing beyond purely technical applications. Yet, a closer look reveals that the students most often represented in these studies are not necessarily the most diverse. Only a small fraction of the 70 studies explicitly included students with disabilities or those who struggle academically. In other words, while AI is often celebrated for its potential to personalize learning, it’s not yet clear whether these technologies are being designed with all students in mind.
Text-Based Learning Still Dominates
We also discovered that the majority of AI systems used rely heavily on text. Chatbots and intelligent tutoring systems remain the most common types of tools, and most student–AI interactions happen through written input and output. For many students, text-based systems are accessible and familiar, but for others, they can pose barriers. Students with dyslexia, speech-language difficulties, or limited literacy proficiency may struggle to engage meaningfully when AI tools only “speak” one language: text. This finding highlights a significant gap between technological advancements and inclusive design. Even as AI becomes more sophisticated with voice recognition, image analysis, and multimodal generative tools, classroom applications often fall short in terms of accessibility and user diversity.
A Missed Opportunity for Multimodal, Human-Centered AI
One of the most promising directions for education lies in multimodal AI, which allows students to interact using voice, images, gestures, or even physical movement. Such systems can adapt to a variety of sensory and cognitive strengths, providing alternative ways for students to engage, express understanding, and receive feedback. However, we found that multimodal approaches were far more common in AI-to-student output (for example, when an AI delivers text plus images or audio) than in student-to-AI input. In simpler terms, the AI may “talk” in multiple ways, but students are still expected to “respond” in just one. This imbalance reveals both a technological limitation and an untapped opportunity for more equitable, human-centered design.
Why It Matters for Teachers
AI holds great promise for making education more adaptive and efficient, but efficiency isn’t the same as fair access. If the tools of the future are designed and implemented without considering learner variability, we risk amplifying the very gaps we hope to close. For educators, this means looking beyond surface-level adoption of AI tools. It’s not enough to integrate AI into the curriculum—we must ask who benefits, who might be excluded, and how we can ensure accessibility for all. This also requires grounding AI integration in learning theories and evidence-based practices that enhance learning for all students. Our review highlights a need for more collaboration between educators and technologists. By working together across disciplines, we can ensure that AI tools are not only powerful but also pedagogically sound and effective.
Looking Ahead
Generative AI has recently transformed how educators and students think about teaching and learning, yet our findings suggest that K–12 applications remain cautious and limited. This caution is understandable—AI’s influence on the cognitive and motivational development of younger students is still an open question. Future research must explore how to balance innovation with responsibility, ensuring that AI supports rather than replaces the essential cognitive work of learning. Our scoping review is just the starting point for this conversation. It offers a bird’s-eye view of how AI is currently being used and, more importantly, where the field needs to go next. If we want AI to personalize learning truly, we must design systems that listen and respond to the diverse ways students learn, think, and express themselves. Only then can AI become not just a classroom tool, but a partner in creating more empowering learning environments for all students.
Read the full article in the Journal of Computer Assisted Learning:
Seung, Y., Basham, J. D., Kim, T., & Lohoefener, J. (2025). Exploring Artificial Intelligence Integration and Student–AI Interaction in K–12 Education: A Scoping Review. https://doi.org/10.1111/jcal.70144
If you have any questions related to this article, don't hesitate to get in touch with Yerin Seung at yerinseung@ku.edu.
Join the conversation in our community!
CIDDL is committed to providing high-quality resources to support the increasing knowledge, adoption, and use of a range of educational technologies that can be used for educators, related services, or leadership preparation programs. For more resources, including videos and blogs, subscribe to our newsletter and follow us on YouTube, Facebook, and LinkedIn. The most important part of our CIDDL community is YOU. Join our community and share the innovative ways you are using technology, ask a question about technology integration, or participate in our bi-weekly live AI Community Chats. We look forward to seeing you in our community!