Module 3: OpenAI and ChatGPT

Module Overview

This module introduces you to ChatGPT and the revolutionary impact of large language models on natural language understanding and generation. You'll explore how these powerful AI systems work, learn the art and science of prompt engineering to effectively communicate with AI models, and develop a comprehensive understanding of the biases and limitations inherent in current LLM technology.

Through hands-on experience with ChatGPT, you'll gain practical skills in leveraging AI for creative problem-solving while developing a critical awareness of the ethical considerations and technical constraints that shape these transformative technologies.

Learning Objectives

  • Introducing ChatGPT: A Revolution in Natural Language Understanding
  • Prompt Engineering: The Art and Science of Guiding AI Models
  • LLM Bias and Limitations: A Comprehensive Analysis

Guided Project

This guided project does not have a traditional repository or notebook materials. Instead, you'll engage directly with ChatGPT through its web interface. For students interested in exploring additional technical background, you can review the legacy AutoEncoders material as supplementary content, though the current guided project and assignment are the primary focus.

OpenAI's ChatGPT

Module Assignment

This module features a unique assignment format that differs from our typical technical implementations.

Brainstorming Innovative Projects with ChatGPT

Objective:

The primary goal of this assignment is to brainstorm innovative project ideas that leverage the capabilities of modern Large Language Models (LLMs) like GPT-4. By using the ChatGPT website, you will engage with the model to come up with at least three compelling project ideas that harness the potential of these advanced language models in various domains.

Prerequisites:

  • An internet connection to access the ChatGPT website.
  • A method to document your ideas (Google Docs, Markdown, etc.)

Steps:

  1. Get Familiar with ChatGPT
    • Go to the ChatGPT website and engage in some initial interactions to understand its capabilities and limitations.
  2. Select a Domain
    • Decide on a general domain you are interested in for your project (e.g., healthcare, sustainability, automation).
  3. Brainstorming Session
    • Use ChatGPT to brainstorm potential LLM project ideas within your chosen domain.
    • Pose questions to the model like, "What are some unique ways to use AI in healthcare?" or "How could machine learning improve public transportation?"
    • Take notes during the brainstorming session for later review.
  4. Idea Refinement
    • Choose at least three project ideas that interest you.
    • Engage with ChatGPT to explore these ideas in more depth.
    • e.g., "What would be the steps to use AI for real-time language translation in emergency services?"
  5. Documentation
    • Document each of your selected ideas, outlining:
      • The problem the project aims to solve
      • How modern LLMs like ChatGPT could be utilized in the solution
      • The data and resources you think you'll need
      • Limitations and challenges of the idea
  6. Peer Review (stretch goal)
    • Share your documented ideas with classmates or a friend for a review and gather constructive feedback.
  7. Final Refinement
    • Refine your ideas based on the peer feedback received.
    • Make necessary adjustments before the final submission.
  8. Submission
    • Submit your well-documented project ideas for evaluation.

Evaluation Criteria:

  • Innovation and creativity
  • Feasibility of the idea
  • Clarity and depth in documentation
  • Peer review feedback (optional)

Resources:

  • ChatGPT website
  • Articles and research papers on the capabilities and limitations of LLMs
  • Examples of projects that utilize LLMs

Additional Resources

ChatGPT and OpenAI

Prompt Engineering

LLM Limitations and Ethics