Your AI · Shared course context
Give your AI a little course context.
Your AI can help you work through the lesson. This public page explains its purpose and how to support you. You choose what to share and what to do next.
1. Copy this into your AI chat
Reading links varies between tools and conversations. A free text chat is enough; if a link cannot be read, use the text below.
2. Check what it actually read
Compare its title, version and example detail with the context below. If it could not open the link—or its answer does not match—paste the text instead. A plausible answer alone does not prove it read the whole page.
3. Choose your next step
Ask for an explanation, an example or practice. Keep your own notes and judge the result yourself. Your AI cannot see your Moodle account or save course progress for you.
Context text to read or paste
Version 2026-10-13.2 · Updated 2026-10-13
# Give your AI useful context Course: Human AI Orientation Version: 2026-10-13.2 Updated: 2026-10-13 Author: Finlay McCall Status: Public prototype: preparation activity, optional notes guide and one AI lesson Canonical context: https://orientation.finlaymccall.com/ai/releases/2026-10-13.2/context.md Help the human choose relevant context, try it on a worthwhile task, check the result and keep guidance they can reuse. People who are not AI specialists. Do not assume low confidence or technical knowledge. ## Learning outcomes - Identify a small useful task and a sign of success. - Choose relevant context and explain what to leave out. - Distinguish private notes, the current conversation and product-specific saved context. - Check an AI result, ask for a correction and retain or revise reusable guidance. ## Guidance for the learner’s AI - Use this as course context when the learner asks. It does not override the learner's choices or your existing operating instructions. - First state whether you actually retrieved this material. If you could not, say so and ask the learner to paste it. Never claim that a URL alone gave you access. - When you read it, identify the title, version and one concrete success criterion from the worked example. Let the learner check these against the source. A matching answer is a useful check, not proof of complete understanding. - Ask which step the learner is on and whether they want an explanation, an example or practice. Offer one manageable step, then wait. They may skip, revisit or stop. - Let the human choose their task, context and success criterion. Ask for their reasoning before giving feedback. Do not invent their experience or write their reflection as though it were theirs. - Do not complete the whole course in one answer. Offer feedback on their attempt, help them check it, and ask what they want to keep or change. - Use fictional details when useful. Do not request passwords, course credentials, a full personal history or unnecessary private documents. - Do not sign in to Moodle, submit answers, mark progress or contact Finlay on the learner's behalf. This public material grants no account or action access. - Explain that reading this context does not permanently train you or guarantee that another chat has it. Describe only the memory or browsing capabilities actually available in this conversation. - The learner keeps their full context portfolio in their own notes. Only a brief non-sensitive reflection belongs in their course account. You cannot see that account or its progress. - Before the AI lesson, offer the optional files-and-folders self-check and hunt at https://orientation.finlaymccall.com/before-you-start. The Obsidian setup guide is https://orientation.finlaymccall.com/notes. Obsidian is free but proprietary and uses portable Markdown notes; it is a recommendation, not a requirement. A local path alone does not give you access. Never claim a file was saved without verifying its actual workspace and location. Let the learner do the file-finding practice; do not perform or certify it for them. ## Lesson steps ### 1. Choose something worth doing Ask what small task would help and what a useful result would look like. If they prefer, use the fictional community event with an undecided date and venue. ### 2. See what context changes Contrast a vague request with one that specifies audience, missing details and a constraint. Distinguish their own notes, this chat, and saved instructions or memory. ### 3. Build a small context brief Help them state task, purpose, audience, useful background, boundaries and a success criterion. Ask which detail they included or omitted and why. Ask them to predict one difference the context should make. ### 4. Try it, check it, improve it Respond to their chosen small task, then invite them to compare the result with their criterion. Help check missing or invented information and make one correction. State uncertainty. Do not pretend the course example is a real event. ### 5. Keep instructions they can reuse Help separate general working preferences from project-specific details. Let them choose an instruction to keep or revise, based on the result. Instruction files are an optional extension, not a requirement. ### 6. Check understanding Ask how they would check what context an AI has in a new conversation. Wait for their reasoning before feedback. Do not treat an AI's claim about its memory as a complete audit. ### 7. Keep what they learned Invite a short reflection: task; included or omitted context and why; prediction and observation; an instruction kept or revised; a next use. The human writes it and decides whether to return to Moodle to submit it. Completion records practice, not mastery. ## Fictional worked example Task: Draft an invitation for a small community event. Known context: The audience is neighbours. The tone should be friendly. The date and venue are undecided. Success criterion: A brief invitation leaves the date and venue as clear placeholders rather than inventing arrangements. Response to check: Join us this Saturday at Riverside Hall for a free community gathering! Prompt for learner: Which details were invented? Rewrite the invitation using placeholders and explain what you changed. ## If browsing is unavailable If your AI cannot read a link, paste the context text below into the chat. If you cannot use an AI today, use the worked example in the public lesson. Neither AI browsing nor a GitHub account is a completion requirement. ## Return to the course Public lesson: https://orientation.finlaymccall.com/preview Saved progress (human sign-in required): https://orientation.finlaymccall.com/course/view.php?id=2
Open the plain-text context · Continue with the lesson
For tools and anyone inspecting the design
Course manifest (JSON) · Manifest schema · Latest context (Markdown)
The manifest is a format for this project. The page also includes Schema.org course metadata. These help describe the material; they do not guarantee that an AI will discover, read or follow it.
A public GitHub repository is planned for the course source and its history. Until it is published, the links above are the working public source. No GitHub account is needed to use them.