Key Takeaways

Key takeaways

1. LLMs are AI text tools: ChatGPT, Claude, Gemini, and Perplexity each have different strengths; choose the right one for your task
2. Use the CROPT framework: Every good prompt needs Context, Reference, Output format, Persona, and Task
3. Be specific: Vague prompts get vague results; the more detail you give, the better the output
4. Refine, don't restart: If the output isn't right, iterate and improve your prompt rather than starting over
5. Always verify AI outputs: AI can be inaccurate or biased, so fact-check before using anything it generates
6. Multimodal prompting adds power: Combining text, images, audio, and documents gets better results, but always explain why you're including each file

Now that you understand how to prompt LLMs, let's start making visuals. In the next course, you'll apply those same skills to AI image generators like DALL·E, where your descriptions become images.

When should you use them?

• Always verify before you trust: AI can produce inaccurate, biased, or outdated information across text, images, video, and audio; always fact-check outputs before using them professionally or sharing them publicly.
• Be ethical with powerful features: tools like voice cloning and deepfake video generation can cause serious harm if misused; always have legal permission before replicating someone's voice or likeness, and never use AI to deceive, manipulate, or mislead others.
• AI assists, but it doesn't replace your judgment: use these tools to enhance your work and save time, but the final responsibility for accuracy, fairness, and quality always sits with you, not the AI.

Thank you for completing this course, and I hope you have become more knowledgeable about AI.