An AI-powered agent combines traditional algorithmic behavior with Large Language Models (LLMs) to create more sophisticated, context-aware interactions. Unlike rule-based agents that follow predetermined logic, AI agents can understand natural language, learn from interactions, and generate dynamic responses.

Brief:

Design and implement an AI-powered agent (personality quiz/ digital persona profiler) using p5.js and an LLM API (like OpenAI's GPT) that can engage in meaningful interactions with users.

Requirements:

Create an AI personality quiz with p5.js and LLM Integration

• Integrate with an LLM API (e.g., OpenAI, Claude, or similar)
• Design a clear interaction flow between user input, LLM processing, and agent response
• Create meaningful prompts that guide the LLM's behavior
• At least four different types of personalities that people may get sorted into.

Design Natural Interaction Patterns

• Implement at least three questions to identify your personality through the interaction with the agent:
  • Text-based communication
  • GUI (buttons, slider, touch)
  • Voice input/output (optional)
  • Image input (optional)
• The results of Personality analysis (minimum 4 outputs) must be consistent with the same sets of the answers selected.
• Design clear interaction flows and recovery paths

Create a System Diagram

• Document the flow of interactions between User, LLM and other components
• Show the relationship between frontend, API, and LLM

Submission:

• Live demo in class (Apr 11, 2025)
• URL to working p5.js implementation