Overview

“Re:Imagine Safe Space” invites users to slow down, reflect, and emotionally reconnect by co-creating personalized AI-generated environments. Through RFID-tagged tactile objects and Gemini Flash 2.0, users trigger ambient scenes that evolve based on symbolic choices—building their own "safe space" through play, memory, and emotion.

This project explores how interactive design and AI can support mental health and introspection in a fast-moving world, transforming fleeting comfort into an intentional, co-created experience.

‍

Context

In fast-paced environments, people often lack dedicated spaces to pause, reflect, and process emotions. While the concept of "safe space" exists, many may not recognize their own, let alone intentionally create one. This project explores how AI and interactive design can facilitate a journey between the mental and physical realm to help users define their own sense of safety and comfort.

‍

Process

Highlights

• Studied therapeutic frameworks to design an emotionally intuitive journey.
• Conducted 12 user interviews and A/B testing of interaction models (3D objects, abstract icons, color symbols).
• Developed a non-directive AI dialogue system, avoiding prescriptive results and encouraging self-discovery.
• Built with a fully custom Arduino + RFID setup, integrated with Gemini AI for ambient generation.

‍

Designing the Safe Space Journey

• Studied therapy session structures to map emotional engagement.
• Created a user journey flow that mirrors reflective processes found in therapy while keeping it exploratory.
• Developed AI dialogue prompts that encourage self-discovery without leading users to specific answers.

Physical & AI Interaction Design

• Designed a system where users select objects representing their personal feelings of safety.
• Integrated AI-generated visuals & soundscapes to dynamically create a personalized safe space based on user input.
• Allowed adjustable sensory parameters (light, movement, sound, density, texture, and openness) for customized experiences.

A/B Testing: Exploring Different Interaction Methods

To understand how users interpret and interact with safe space elements, I conducted an A/B test with three different input methods:

VersionInteraction StyleUser FeedbackA: 3D Physical Objects - Users selected and placed 3D objects that represented their emotions.Felt tactile and engaging, but some users found the objects too limiting in terms of representation.B: 2D Black & White Icons - Users selected icons representing abstract concepts (e.g., waves, trees, doors).More flexible than 3D objects, but some users felt less emotional connection due to lack of visual richness.C: 2D Colorful Icons - Users selected colorful illustrated icons representing emotions and environments.Elicited stronger emotional engagement, but some users felt the colors were too prescriptive and influenced their choices.

A/B Test Findings:

• 3D objects worked well for tactile learners, but felt too static for users who preferred abstract thinking.
• 2D black & white icons allowed for more interpretation, but some users struggled to make emotional connections.
• 2D colorful icons were the most emotionally engaging, but risked subconscious bias due to color psychology.
• Final Decision: Combined 2D abstract icons + adjustable environmental elements for best balance of flexibility and personalization.

‍

User Research & Interviews

💬 Conducted 12 in-depth user interviews to analyze how people perceive, construct, and interact with the concept of a safe space.

Key Findings from Interviews:

• Many participants never consciously thought about their safe space but intuitively recognized certain environments that made them feel at ease.
• Some associated specific objects, colors, or sounds with safety, while others needed dynamic space (motion, breathing elements) to feel at peace.
• Emotional states fluctuate, meaning users wanted a space that could shift with their mood rather than remain static.
• A few users felt uncomfortable with "realistic" AI-generated environments, preferring abstract or dreamlike visuals over literal reconstructions of their past.

Key Design Adjustments Based on Research:

• Allowed users to adjust elements dynamically (light, density, movement, sound).
• Incorporated AI-generated spaces that are slightly abstract, preventing hyperrealism from disrupting immersion.
• Encouraged open-ended reflection rather than leading users toward a predefined idea of safe space.

Outcomes

• Presented at NYC Resistor and ITP Spring Show 2025, drawing strong responses around emotional resonance and accessibility.
• Helped 500+users articulate feelings, connect to memories, and explore emotional safety in a playful, non-threatening way.

What’s Next
I'm looking for new venues—festivals, residencies, or organizations—interested in showcasing experimental tech that prioritizes emotional intelligence, mental health, and immersive storytelling.
🤝 If you're working at the intersection of AI, interaction design, and well-being, or know someone who is—I’d love to connect.

‍

Let’s Collaborate
I'm open to roles and partnerships that explore how AI can foster intimacy, play, and growth—rather than overwhelm. If you’re building something that makes technology feel more human, more healing, or more curious—I’d love to be part of that.