BeautAIful
Teaching Teens Media Literacy in the Age of AI
Overview
BeautAIful is a UX research project designed to evaluate how we can help teens critically interpret AI-altered images and understand how artificial intelligence shapes beauty standards and self-perception. I led UX research and learning experience design to evaluate how AI-focused education could strengthen digital literacy and self-awareness.
​
Team: 2 researchers, 1 professional mentor
Tools: Qualtrics, Figma, ATLAS.ti, Excel
Methods: Mixed-methods research, usability testing, pre/post assessments, qualitative analysis
Context & Challenge
Social media filters and AI image generators (like Stable Diffusion and FaceApp) are redefining what “beauty” looks like online. Teens scroll through AI-enhanced photos daily often without realizing how heavily algorithms influence what they see.
​
Problem:
Teens can navigate technology, but they struggle understand how AI shapes the images and ideals they consume.
​
We set out to research how an an e-learning experience could:
-
Increase AI media literacy
-
Strengthen critical thinking around synthetic media
-
Improve self-perception and digital confidence
My Role
-
Designed the research plan and learning objectives using Bloom’s taxonomy.
-
Led testing sessions with 18 teens (ages 13–17).
-
Designed interactive learning modules that visualized how LLMs and diffusion models alter content.
-
Analyzed both quantitative gains and qualitative feedback to measure impact.
-
Ensured ethical and age-appropriate participation.
Research Goals
Core Question:
How can AI-focused digital literacy education help teens interpret AI-altered media and strengthen self-perception?
​
Learning Objectives:
-
Understand how LLMs and image-generation models create and modify media.
-
Identify signs of AI manipulation in visual content.
-
Recognize how training data bias influences outputs.
-
Reflect on how AI-driven imagery affects self-image.
-
Develop strategies to critically evaluate authenticity online.
Methodology
Mixed-Methods Evaluation
-
Questionnaires measuring understanding of AI and media manipulation before and after learning module completion.
-
Contextual inquiries where we observed engagement, clarity, and comprehension.
-
Interviews collected reflections on emotional impact and media confidence.
-
Academic research on existing studies.
-
Market research on existing products of similar domain.
UX & LxD Principles Applied
-
Introduced AI concepts step-by-step to reduce cognitive load.
-
Explained how LLMs use training data and addressed consent in AI imagery.
-
Incorporated real-time quizzes and reflection prompts.
-
Chose imagery and examples that represented diverse users.​​
Final Product
Interactive eLearning prototype that prioritized learning design fundamentals and effectively introduced target concepts.
Results & Insights
Quantitative Impact
Post-test accuracy increased from 78.7% → 95.4%, a statistically significant gain (p < .05).
​
Key Takeaways
-
Teens were more confident evaluating digital images after the module.
-
AI transparency boosted curiosity rather than fear.
-
Framing AI as a system with design choices helped users think critically, not cynically.
Impact
BeautAIful showed how UX research + learning design can make AI literacy accessible, empowering, and measurable.
It demonstrated that design principles like feedback loops, ethical framing, and iterative testing, can turn abstract AI concepts into tangible understanding.
​​
Reflection
BeautAIful reinforced my passion for designing human-centered AI experiences.
It deepened my understanding of LLMs, generative models, and AI ethics, and taught me how to translate technical complexity into clear, meaningful learning.​​ I also gained expertise in conducting user studies with children.
​
Next Steps
-
Expand to a module on AI-generated text (LLMs and misinformation).
-
Incorporate adaptive learning using AI feedback loops to personalize content.
-
Test cross-cultural perceptions of AI beauty filters.