This two-part workshop is designed to give learners the confidence and skills to get high-quality outputs from AI chatbots including ChatGPT, Copilot, and more. We’ll work through how they work, safe use, key players, use cases, and a deep dive into prompt crafting—with demos and practice throughout. It’s appropriate both for absolute beginners and those with experience using AI chatbots who feel they still have room to improve.
After the two-part workshop, learners will:
Understand strengths and limitations of AI chatbots, and how to mitigate risk at work
Have a strong understanding of all the ways to use AI chatbots in the assistant workflow
Be able to create strong prompts to optimize outputs
Be using AI chatbots on a weekly (if not daily) basis
All sessions will include practical assignments and takeaways to make learning stick. We’ll engage with learners before, between, and after workshops via the ASAP Circle space for this workshop to build community and drive accountability.
Session 1: AI Chatbot Essentials & How to Prompt for Quality Outputs Tuesday, July 16, 2024, 1PM-2:30PM ET
This session covers:
Defining LLMs—how they work, strengths, limitations, and safe use
Comparison of major AI chatbots and paid vs free versions
Fundamentals of prompt crafting and where most people go wrong
Breakout exercises to practice prompting through real-world projects for EAs
Bonus Resource: COIF Prompt Framework, bite-sized homework assignment to do between sessions (10-15 mins)
Session 2: Advanced Prompting Techniques for Complex Projects Wednesday, July 17, 2024, 1PM-2:30PM ET
This session covers:
Advanced prompting (including megaprompting, prompt generators, creating custom GPTs, and more)
Prompt dissection
Breakout exercises to practice more advanced prompting through real-world projects for EAs
Bonus Resource: Prompts Bank
Hot Tip: When you're chatting with ChatGPT (or any AI chatbot), imagine it's an eager and enthusiastic yet really green intern. You need to *really* spell out what you want—give it all the context and even copy-paste in examples you've created yourself so it knows the quality and context of what you’re looking for.