AI IN ACTION / RICHARD’S LIVE DEMOS Shared introduction: slides 8–10. Approximately four minutes each. Use the example files as clearly identified backups. They are authored workshop examples, not saved outputs from Richard’s NotebookLM or Work account. 1. NOTEBOOKLM / PDFs into media Source: https://ai-in-action.stemed.ca/starter_files/Verified_Survey_Facts.pdf Ready infographic: https://ai-in-action.stemed.ca/examples/Survey_Infographic.pdf Before Monday: Create a notebook and add Verified_Survey_Facts.pdf. Generate an infographic and an Audio Overview, then check them. Keep the finished outputs open and download the audio where the account allows it. Do not spend the live slot waiting for both outputs to generate. Live question: Using only this source, what did the 48 unique fictional responses ask AI to help with next? List each task and its count, then cite the source. Explain why these data cannot support a claim about real Ontario educators. Infographic customization: Create a clear, playful infographic for educators titled “What would we make next?” Show the six next-task categories and their exact counts. State that this is synthetic practice data with 48 unique responses, after removing two exact duplicates from 50 raw rows. Do not imply representative findings or measured time savings. Make the most requested task easy to spot. Audio customization: Explain the six next-task categories in a short, friendly conversation for educators. Use only the supplied source. Open with the most requested task, explain one practical implication as a suggestion, and end with the synthetic-data limitation. Avoid claiming that this fictional survey represents real teachers. On screen: Ask the question, open one citation, show the prepared infographic, and play 20–30 seconds of the prepared audio. Check lesson planning = 16 of 48 and all six counts sum to 48. Ask James which representation he would use with learners. If slow: Show the prepared NotebookLM result. If unavailable, show the workshop infographic and identify it as the authored example. It is not a substitute recording of the audio. Google documentation: https://support.google.com/gemininotebook/answer/16206563?hl=en 2. STEM | ED / One real app: WORK | SHOP Open: https://stemed.ca/ and select WORK | SHOP, or use Richard’s bookmarked app and prepared workshop. Before Monday: Pre-open one actual WORK | SHOP workshop Richard can show. Choose one participant-facing workflow that already works. Prepare the starting screen and its completed result; use a fictional or demonstration workshop when appropriate. On screen: Name the teacher problem the workshop solves. Show one action in the real app and the resulting screen or resource. Explain one specific change Richard asked AI to make while building it and the check he used. James connects this to clearer instructions and reusable context. Keep the demonstration to this one application; the full STEM | ED gallery is an optional reference on slide 20. Closing question: “What small tool would make one repeated job easier for you?” If the app cannot load: Use the preserved real STEM | ED screenshot on slide 20 and tell the building story. Do not describe a live action as completed if it did not run. 3. CHATGPT WORK / A Maker Day poster people can open Preferred: Work on this computer or https://chatgpt.com/ with Work selected. Alternative: Give the same source and task to an available dot. Rehearse its file delivery before the session; connected local-computer access is separate from its cloud computer. Source: https://ai-in-action.stemed.ca/starter_files/Maker_Day_Source_Brief.pdf Ready poster: https://ai-in-action.stemed.ca/examples/Maker_Day_Launch_Poster.pdf Starting prompt (attach Maker_Day_Source_Brief.pdf): Create a finished, downloadable, one-page PDF launch poster for this fictional Maker Day. Title it “Make, Test, Share”. Aim it at secondary learners. Make it playful, clean and readable from across a classroom, with a strong visual and lots of breathing room. Include 24 learners, three groups of eight, three stations and the 60-minute duration. Describe Motion Lab, Bridge Build and Design Story in one short sentence each. Preserve the 20 cm bridge span and the source’s activity details. Do not invent a date, school or location. Add a discreet “Fictional workshop example” footer. Open the PDF preview and give me the download. Check these facts against the source before delivering it. Make the actual PDF file, rather than only giving me poster text. Live revision: Keep all facts and the one-page PDF format. Make the headline more playful and give the three station names more visual emphasis. Deliver the revised PDF and tell me what changed. On screen: Submit the prompt, show how the task creates a file, open the completed PDF, compare one source fact and request the revision. Keep a completed run open in another tab. If generation takes longer than 45 seconds, show that run while the current task continues. If no account-generated result is available, identify the supplied poster as the authored workshop example. Check: 24 learners; 3 groups of 8; 3 named stations; 60 minutes; 20 cm span; no invented event logistics. Local option: Give Work the source file’s full path and a dedicated demo output folder. Online option: attach the source PDF and download its result. Both routes use the same task; no app connection is necessary for this file-based demo. OpenAI documentation: https://learn.chatgpt.com/docs/web Files: https://learn.chatgpt.com/docs/artifacts-viewer Dots: https://learn.chatgpt.com/docs/dots/computers-and-apps HANDOFF TO PROJECTS James’s instructions, Richard’s media and finished files, and the real application all use the same pattern: a clear task, relevant context, an output and a check. Explain the workspace around the model, then split into the two groups. Richard’s participants choose ONE project. STEM | ED remains show-and-tell, not a participant project.