AI Resource Duck: Clickable Prototype and Usability Testing

AI Resource Duck: Clickable Prototype and Usability Testing

Introduction

AI Resource Duck is a macOS desktop application designed to help students develop healthier and more mindful relationships with AI tools. As AI assistants like ChatGPT, Claude, and other Copilots become increasingly integrated into academic workflows, many students lack visibility into how much they rely on these tools and whether that reliance supports learning or replaces their own thinking.

AI Resource Duck addresses this challenge through gentle awareness rather than restriction. The application features a small desktop duck that lives on the screen and tracks AI usage across apps and web browsers. As usage increases throughout the day, the duck’s mood gradually shifts—from happy to overwhelmed—providing a playful visual metaphor for cognitive load. A dropdown menu panel provides more detailed statistics about time spent using AI, number of sessions, prompts sent, and breakdowns by AI tool.

Through usability testing, we aim to evaluate three key aspects of the design:

  • Discoverability: Can users easily find and access the main features of the interface?
  • Comprehension: Do users understand what the duck’s mood represents and how their usage affects it?
  • Emotional response: Does the experience feel supportive and reflective rather than judgmental?

Usability Testing Script

https://docs.google.com/document/d/1JNdN0Qkgk678kLBALGBAXQqyjj8LwEXqhTm38EOu83k/edit?usp=sharing

Our usability test follows a task-based format inspired by the “Don’t Make Me Think” approach to usability testing. Participants are asked to interact with the prototype while thinking aloud so we can observe how they interpret the interface and navigate key features.

Participant Profile

Participants are college students who regularly use AI tools such as ChatGPT, Claude, or Copilot for academic work.

Test Structure

Each usability session lasts approximately 20–25 minutes and includes:

  1. Background questions about AI usage habits
  2. First impressions of the duck interface
  3. Task-based exploration of the app features
  4. Reflection on emotional tone and usefulness

Clickable Prototype

DMG Install: https://drive.google.com/file/d/1JA62U7YuZ7mpiHIp7QzVP-ewerPVLfnb/view?usp=drive_link

Our clickable prototype is a working macOS desktop application built using PyQt6 and PyObjC and Sh, allowing users to interact with the system as they would with a real product.

Unlike static mockups, the prototype includes live interactions, animations, and data visualization.

Key Features of the Prototype

Desktop Duck Pet

The central interface element is a small animated duck that lives on the user’s desktop.

The duck’s mood changes based on AI usage:

Usage Level Duck Mood
0–25% Happy
25–50% Concerned
50–75% Stressed
75–100% Overwhelmed

This provides a visual metaphor for cognitive load.

Users can also:

  • Drag the duck
  • Move it around the screen
  • Add multiple ducks
  • Watch mood changes over time

Menu Bar Dashboard

Clicking the duck icon in the macOS menu bar opens a dropdown dashboard displaying:

  • AI usage percentage
  • Time spent using AI
  • Number of sessions
  • Number of prompts sent

These statistics help users better understand their AI habits.

Reflection Prompts

At higher usage levels, the interface displays reflective prompts encouraging users to pause and consider their learning process.

These prompts are intentionally non-judgmental and supportive.

Source Breakdown

Users can expand the Sources panel to see which tools they used most, such as:

  • ChatGPT
  • Claude
  • Cursor
  • Copilot

This helps users understand where their AI time is being spent.

Daily Goal Settings

Users can set personal goals for:

  • Maximum AI minutes per day
  • Maximum sessions
  • Maximum prompts

The system does not enforce hard limits; it simply provides awareness and feedback.

Technical Implementation

The prototype was built using:

  • PyQt6 for the user interface
  • PyObjC for macOS integration
  • SQLite for local data storage
  • AppleScript polling to detect active applications
  • Custom animation systems for the duck pet

This allowed us to simulate a realistic interactive experience for usability testing.

Prototype Limitations

Current limitations include:

  • macOS-only compatibility
  • Limited browser tracking
  • Local-only data storage

Future versions could include:

  • Mobile companion apps
  • Weekly analytics dashboards
  • Chrome extension integration
  • More pets!

Conclusion

AI Resource Duck explores a new approach to AI literacy by focusing on awareness rather than restriction. Instead of blocking AI usage, the design encourages students to reflect on how they use AI through playful interaction and clear usage feedback.

Our usability testing will help determine whether the “duck metaphor” effectively communicates AI usage patterns and whether the interface encourages mindful reflection without creating feelings of guilt or surveillance.

Avatar

About the author