This is my first hands-on reflection project from the Udacity Generative AI Nanodegree. In this activity, I explored the behavior, capabilities, and limitations of a commercial Large Language Model (LLM), specifically ChatGPT by OpenAI.

To understand how LLMs like ChatGPT interpret prompts, generate responses, handle follow-up context, and where they might fail.

What is Generative AI and what are its applications?

• Defined Generative AI as models that can create text, images, code, audio, and more using learned patterns from data.
• Explained training using LLMs and GANs.
• Listed key applications:
  • Text generation (blogs, translations, summaries)
  • Image/art generation (AI illustrators, avatars)
  • Audio/video (TTS, AI music, lip-sync)
  • Code generation (Copilot, learning tools)
  • Science & medicine (drug discovery, simulations)
  • Business productivity (reports, assistants)
  • Gaming (NPC dialogue, story design)

“It gave me a more detailed and broader overview than expected. I discovered new use cases I wasn’t even aware of—especially in science, gaming, and productivity.”

For the first of these applications, how do we ensure the ethical use of Generative AI?

• Provided 8 key ethical guidelines:
  1. Content moderation & toxicity filtering
  2. Bias mitigation via audits & diverse datasets
  3. Transparency & disclosure of AI involvement
  4. Fact-checking using RAG or APIs
  5. Privacy & consent for training data
  6. Human-in-the-loop for high-stakes outputs
  7. Purpose restrictions (e.g., no full-essay cheating)
  8. Explainability (citations, traceability)

“It was able to continue the conversation smoothly and remembered the context from the first question. The ethical framework it listed was surprisingly structured and practical.”

Can you find a question the LLM cannot answer well?

“I tried to confuse it with logical traps, but the LLM handled them fairly well. It’s clear that these models are getting smarter — and harder to ‘stump’ on basic reasoning.”

exploring-generative-ai-llm/ ├── images/ │ ├── llm_response_applications.png │ └── llm_response_ethics.png ├── reflections/ │ └── my_thoughts.txt ├── README.md 

• Prompt Engineering
• LLM Evaluation
• Ethical AI Awareness
• Contextual Reasoning
• OpenAI (ChatGPT)

• LLMs are excellent at understanding prompt context and generating detailed outputs.
• Ethical considerations must evolve alongside AI capabilities.
• Even basic prompts can lead to rich learning if explored deeply.

🔄 Inspired by the Udacity Generative AI Nanodegree 📩 Feel free to reach out if you’d like to collaborate on real-world LLM projects!