Human-Centered AI Design for Learning & Growth: Design Principles, Framework, and Application

Overview

As generative AI becomes a central tool in learning and productivity, I've been observing that when not intentionally designed, AI can reduce human critical thinking and discourage reflection. Many tools today optimize for speed and convenience, but often at the cost of user agency and long-term growth.

In response, I explored how to design AI that supports learning, not just answers. That amplifies human strengths, not replace them.

Based on the previous academic research projects (2021-2025) done in the field, observation, and practical application, I developed:

• Four design principles for human-centered AI that promotes growth and agency,
• A three-phase framework inspired by self-regulated learning theory, and
• Wireframes and a front-end level, conceptual prototype that applies these ideas to an AI-powered writing assistant.

Together, these elements propose a more intentional, human-centered direction for designing AI systems that help people think better, and become better.

Design Principles

Rooted in HAI (Human-AI Interaction) and education research, I derived five design principles for human-centered AI that fosters learning.

Framework

Inspired by self-regulated learning (SRL) theory¹ and the design principles, I then developed a framework to propose a 3-phase structure that could be applied to AI systems to support human growth. In each phase, AI plays a distinct pedagogical role, not to give answers, but to guide, scaffold, and spark critical thinking.

Putting into Action: Wireframe

Here’s how this framework could be realized in an AI-powered learning assistant. I focused on one of the most popular use cases these days, academic writing, which is critical for students in higher education.

Current Status

This project is currently being developed into a functional web-based prototype.

As a designer working in the age of generative AI, I'm working on rethinking the design process itself. I’ve been exploring how to leverage multimodal Gen AI tools and "vibe-coding" methods to accelerate prototyping without compromising depth or clarity.

So far, I’ve found that:

• Well-structured wireframes paired with prompt-based AI tools significantly reduce UI design to prototyping time.
• My ability to read, edit, and organize front-end code lets me quickly transform concepts into high-fidelity, engineer-friendly prototypes.

By building this tool as both a product and a methodological experiment, I'm aiming to develop faster, smarter, and more collaborative design workflows tailored for AI-enhanced futures.

Reflection (so far...)

This project was driven by my vision to design technologies that amplify human potential through learning.

I realized that human-centered AI is not just about usability or productivity, but rather about shaping the future of how we grow, think, and interact, and designers have a responsibility to guide how AI influences human behavior.

I aim for this ongoing project to reflect my core design philosophy and an opportunity to explore the kind of work I want to pursue moving forward.

References

1. Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview.
2. Steinert, S., et al. (2023). Harnessing Large Language Models to Enhance Self-Regulated Learning via Formative Feedback.
3. Chang, D. H., et al. (2023). Educational Design Principles of Using AI Chatbot That Supports Self-Regulated Learning in Education: Goal Setting, Feedback, and Personalization.