Skip to main content

Final Reflection

Final Reflection

Designing a campus companion with social responsibility

Campass began as a playful response to a practical problem: students often experience campus as disconnected buildings, routes, and tasks. By combining a simplified map, NFC check-ins, AR mascot encounters, a message wall, and a collection loop, the design reframes orientation as an active social journey.

1. Social Implications

The strongest social value of Campass is its potential to turn campus exploration into a shared experience. Location-based posts and collectible achievements can make unfamiliar spaces feel warmer and more approachable, especially for newcomers who may otherwise experience orientation as stressful or isolating.

Belonging

Shared discoveries and place-based posts help students see campus as a living community rather than a static map.

Local Knowledge

Informal tips, stories, and hidden details make campus knowledge easier to pass between cohorts.

Motivation

Gamified progress gives users a reason to revisit spaces and notice details they may otherwise ignore.

A public message wall can become noisy, exclusionary, or harmful if it lacks moderation. Gamification may also pressure users to visit every location or compare progress with others. Campass should therefore keep social participation optional, lightweight, and respectful of different comfort levels.

2. Ethical Considerations

Because Campass is built around location-based interaction, privacy is the most important ethical concern. NFC check-ins and AR encounters can create meaningful engagement, but they also imply that the system may know where a user has been and what they have unlocked.

Collect only the data needed for the experience.
Explain clearly what is stored and why.
Keep public posting optional rather than mandatory.
Provide moderation for community-generated content.
Offer non-AR and accessible alternatives for key interactions.

3. Use of Generative AI

Generative AI was used as a supporting tool, not as a replacement for design judgment, user research, or technical validation. Each tool had a clear role in the project workflow.

Hunyuan3D

AR model generation

Used for generating and supporting the 3D mascot/AR asset workflow. Final assets still required human selection, integration, and testing in the prototype.

VS Code + Copilot + Codex

Code implementation

Used for debugging, component edits, and documentation. Adhered to "Vibe Coding" workflow with manual verification of every component. Detailed logs available in AI Log.

View our AI Logs

Gemini

General ideation and writing

Used for brainstorming, wording refinement, and structuring explanations. Final claims were grounded in our project decisions and prototype evidence.

Nano banana + GPT image 2

Visual & Icon generation

Used for visual generation support where initial assets were needed. Generated visuals were treated as design materials rather than factual evidence.

The main limitation of using AI is that it can produce confident but generic statements. To avoid this, we grounded final decisions in the actual Campass concept: NFC-based check-ins, AR mascot summoning, the message wall, collection progress, and the performance constraints documented in our implementation.

4. Final Takeaway

The most valuable version of Campass is not the most feature-heavy one. It is the version that helps users feel oriented, curious, and connected while still respecting their privacy, attention, and autonomy.

Reference List

  1. [1] Hunyuan3D, v2.0, accessed on 2026-04-10, available at https://3d.hunyuan.tencent.com. Used for generating and supporting the 3D mascot/AR asset workflow.
  2. [2] Nano Banana, vGemini 2.5 Flash Image, accessed on 2026-04-20, available at https://gemini.google.com. Used for image generation of concept visuals and supporting assets used in documentation and poster presentation.
  3. [3] ChatGPT Image 2, v1.0, accessed on 2026-04-10, available at https://chatgpt.com. Used for image generation support and quick visual variation drafting during iterative design communication.
  4. [4] Gemini, v3.1 Pro, accessed on 2026-04-11, available at https://gemini.google.com. Used for higher-depth ideation, structure refinement, and polishing reflection wording.
  5. [5] Gemini, v3 Flash, accessed on 2026-05-02, available at https://gemini.google.com. Used for rapid prompt iteration, quick rewriting, and short-cycle wording checks during documentation updates.
  6. [6] Grok Code Fast 1, v1.0, accessed on 2026-04-10, available at https://x.com/i/grok. Used for fast coding assistance, patch drafting, and debugging alternatives in implementation tasks.
  7. [7] GitHub Copilot, v web version, accessed on 2026-04-10, available at https://github.com/features/copilot. Used for code completion, inline suggestions, and implementation acceleration during development.
  8. [8] Codex, v5.2, accessed on 2026-04-05, available at https://chatgpt.com/codex. Used for early-stage component scaffolding and baseline implementation support.
  9. [9] Codex, v5.3, accessed on 2026-04-26, available at https://chatgpt.com/codex. Used for iterative bug fixes and task-level refactoring in the development workflow.
  10. [10] Codex, v5.4, accessed on 2026-04-22, available at https://chatgpt.com/codex. Used for stronger TypeScript and UI logic cleanup across modular components.
  11. [11] Codex, v5.5, accessed on 2026-05-01, available at https://chatgpt.com/codex. Used for final-round integration checks, consistency updates, and documentation refinement.