AI Learning Hub › Events › Past events › Session Two: Breakout Room Lightning Workshops
Past session · 20 August 2025
Session Two: Breakout Room Lightning Workshops
Seven rooms, forty-five minutes each, one AI workflow per room. Every room’s instructions, files, and trained chat are below, so any of them can be run again — alone, or with a few colleagues and somebody keeping time.
- Presented
- Wednesday, 20 August 2025, 11:30am PT, on Zoom
- Facilitators
- Taryn Marks, Will Huggins, Sergio Stone, Kay Cadena, Heather Joy, Grace Lo, Qi Huang, and Nhatrang Nguyen — one library representative per room
- Format
- Seven parallel breakout rooms, each with a leader, an instructions monitor, a timekeeper, and testers
Written for August 2025. Several rooms assume a paid ChatGPT account, which was the state of the tools then; some of what needed a subscription in these instructions no longer does, and the interfaces have moved. The workflows still hold.
What the session was for
Putting it into practice
- Humans stay in charge — AI never works alone
- You are the senior editor, and you hold final approval
- Bring your own context and expertise to guide the work
- Use the right tool — AI is one option, not always the answer
Learning by doing
- Turn curiosity into action: try the tools on something that actually matters to you.
- Start small, finish something, reflect, improve. Then raise the bar.
- Notice the pattern — goal, context, example, feedback, revision — and repeat it until the result sings.
- Carry the wins forward. Each round builds judgment, speed, and imagination.
Keeping going afterward
- Open the trained chat for a room, add your own materials, and keep working in it.
- Follow the text instructions and the Drive folder for the labs and templates.
- Save what works. Keep a personal prompt library and notes on what each one produced.
The seven rooms
Each room has the same shape: a leader, someone watching the instructions, a timekeeper for the forty-five minutes, and testers giving feedback. The links are the originals.
Room one · Facilitator: Sergio Stone
Creating a writing partner
Tools: ChatGPT (paid), Perplexity (free), and a few pieces of your own writing to test on.
You will:
- Generate AI-driven diagnostics of your writing habits
- Gather authoritative style references to prompt with
- Build, test, and refine a custom GPT writing partner
- Model high-impact writing to eliminate the common AI weaknesses
This room grew into the Writing Partner agent guide and the writing skills, which are the maintained versions of it.
Room two · Facilitator: Heather Joy
Lesson planning with AI
Tools: ChatGPT (at least one paid account in the room) and Gemini.
You will:
- Run citation-rich research on a law-related topic
- Build custom GPTs for lesson plans, rubrics, and assessments
- Apply editing and ethical guardrails to what the AI produced
- Create and share an AI study partner for students
Room three · Facilitator: Kay Cadena
NotebookLM 101
Tools: ChatGPT (paid) and NotebookLM (free).
You will:
- Set up a notebook and populate it with AI-generated sources
- Explore its collaboration, multimedia, and chat features
- Identify and reflect on the use cases that apply to legal work
For the deeper version of this, see Identifying trends in a corpus of documents.
Room four · Facilitator: Qi Huang
AI and datasets
Tools: Google Colab (opens in a new tab) (free) and Google Sheets.
You will:
- Use Colab to clean data, generate visualizations, and test models
- Build interactive tools to synthesize and share a dataset
- Troubleshoot and refine the code the AI wrote for you
Room five · Facilitator: Will Huggins
Deep research and building a presentation
Tools: ChatGPT (paid), Gemini, Perplexity Pro, Gamma, and PopAI.
You will:
- Configure and launch several AI research tools on one question
- Extract the legal insights and auto-generate a deck from them
- Review, refine, and ethically check what came out
Room six · Facilitator: Grace Lo
AI and legal research
Tools: Westlaw, Lexis, and Bloomberg Law accounts, plus NotebookLM.
You will:
- Use the AI features inside Westlaw, Lexis, and Bloomberg Law
- Run a multi-jurisdictional survey using grounded retrieval and NotebookLM
- Compare, analyze, and refine the results from each platform
The follow-up session on reading those tools’ output critically is Assessing output in CoCounsel and Protégé.
Room seven · Facilitator: Taryn Marks
Choose your own adventure: GenAI 101
The room for anyone who was not sure where to start, or who wanted to talk about AI generally rather than build something. No tools required, no prerequisites, and the discussion went where the questions went.
The questions on the table:
- How do I know when to use generative AI — and when not to?
- What does critical thinking look like alongside these tools?
- What is Perplexity, and how is it different from Google?
- What process should I follow to decide whether to use AI at all?
That last question became the PAUSE Rule, and this room’s ground is now covered by Introduction to AI for Beginners.
Privacy and confidentiality. These rooms have you upload your own work to third-party tools. Check how each one treats your data first, keep confidential and client material out, and apply the PAUSE Rule before you start.
Want to run one of these with your team? Email library@law.stanford.edu and we will help you set it up. All past sessions are in the archive.