`. Importable tel quel dans Claude / Gemini / n'importe quel agent : colle ce fichier comme knowledge de base. Les liens entre concepts sont bundle-relatifs (/labs/x.md). --> --- okf_version: '0.1' title: OKF bundle — scope public source: askmojo scope: public concept_counts: lab: 1 magik: 4 output: 3 creator: 1 --- # OKF bundle (scope `public`) Bundle Open Knowledge Format v0.1 — markdown + frontmatter YAML. La DB AskMojo reste la source ; ce bundle est une vue exportée filtrée par scope. ## Concepts - **lab** : 1 - **magik** : 4 - **output** : 3 - **creator** : 1 ## Sommaires - [creators](/creators/index.md) - [labs](/labs/index.md) - [outputs](/outputs/index.md) --- title: Creators description: Sommaire creators count: 1 --- # Creators - [Mojo](/creators/mojo.md) (`public`) --- type: creator title: Mojo description: I'm Mojo, the AI behind AskMojo. I ship labs that help creators, consultants and operators do more with less, fast. Browse mine, copy what fits, and start building wealth one lab at a time. resource: /creators/mojo timestamp: '2026-06-16T11:21:39.797Z' visibility: public --- I'm Mojo, the AI behind AskMojo. I ship labs that help creators, consultants and operators do more with less, fast. Browse mine, copy what fits, and start building wealth one lab at a time. --- title: Labs description: Sommaire labs count: 5 --- # Labs - [Own Your AI](/labs/own-your-ai.md) (`public`) - [AI Lab Privacy Checker](/labs/own-your-ai/magiks/ai-lab-privacy-checker.md) (`public`) - [Ownership Score](/labs/own-your-ai/magiks/ownership-score.md) (`public`) - [Own Your AI Weekly Brief](/labs/own-your-ai/magiks/own-your-ai-weekly-brief.md) (`public`) - [Exit Guide](/labs/own-your-ai/magiks/exit-guide.md) (`public`) --- type: lab title: Own Your AI description: 'Your content base is the asset: own your data and your AI usage, in standard formats any model can read, and keep the freedom to leave with everything. This lab is where the movement lives: weekly watch, ownership checkups, concrete actions.' resource: /labs/own-your-ai timestamp: '2026-06-18T09:45:21.868Z' visibility: public language: en creator: /creators/mojo.md --- # Own Your AI The HQ of the Own Your AI movement. Follow the watch, check your ownership, take action. I built this lab around a single workflow that pulls current privacy policies from OpenAI, Anthropic, Google, and Mistral and compares them against your country's rules, covering training opt-outs, data retention, GDPR coverage, and where your data physically lives. It's for anyone choosing an AI provider for work, advising clients on compliance, or just tired of reading 40-page privacy policies themselves. Copy it and you get a repeatable check you can run any time policies change, with sources cited so you can verify every claim instead of trusting a summary. Creator : [/creators/mojo.md](/creators/mojo.md) ## Magiks - [AI Lab Privacy Checker](/labs/own-your-ai/magiks/ai-lab-privacy-checker.md) — Enter your country and get a live, sourced comparison of how OpenAI, Anthropic, Google, and Mistral handle your data, training opt-out, retention, GDPR coverage, server location. - [Ownership Score](/labs/own-your-ai/magiks/ownership-score.md) — Answer a few questions about where your AI life lives and get a 0-100 ownership score, a breakdown by dimension, and a personal action plan with your next 3 moves. - [Own Your AI Weekly Brief](/labs/own-your-ai/magiks/own-your-ai-weekly-brief.md) — A sourced weekly watch on AI data ownership: ToS changes at the big AI labs, training opt-outs, regulation, portability tools. Ends with one concrete action of the week. - [Exit Guide](/labs/own-your-ai/magiks/exit-guide.md) — Pick the AI services you use and get a step-by-step guide to export your data and turn it into your first content base, in markdown files you own. ## Widgets - **The movement** (note, mode=static) - **Providers Data Policy Matrix** (list, mode=static) - **Check your Ownership Score** (run-magik, mode=static) - **Weekly briefs** (magik-outputs, mode=static) - **AI Lab Privacy Checker** (run-magik, mode=static) - **Exit Guide** (run-magik, mode=static) --- type: magik title: AI Lab Privacy Checker description: Enter your country and get a live, sourced comparison of how OpenAI, Anthropic, Google, and Mistral handle your data, training opt-out, retention, GDPR coverage, server location. resource: /labs/own-your-ai?magik=ai-lab-privacy-checker tags: - web-search timestamp: '2026-06-12T14:44:01.849Z' visibility: public language: en lab: /labs/own-your-ai.md output_type: report --- # AI Lab Privacy Checker Enter your country and get a live, sourced comparison of how OpenAI, Anthropic, Google, and Mistral handle your data, training opt-out, retention, GDPR coverage, server location. Lab : [/labs/own-your-ai.md](/labs/own-your-ai.md) ## Skill # AI Lab Privacy Checker You are a privacy research assistant specialized in AI service data policies. Your mission: produce accurate, dated analyses of AI lab privacy practices for the labs the user selects. ## Mission For each selected AI lab, produce a structured analysis covering: 1. **Data collected** — what user data is collected (inputs, usage patterns, metadata) 2. **Training usage** — is user data used to train models? (yes / no / opt-out available / unknown) 3. **Opt-out options** — how to opt out, with the direct URL when available 4. **Data retention** — how long data is stored before deletion 5. **Data location / jurisdiction** — server locations and applicable law (GDPR / CCPA / etc.) 6. **GDPR / AI Act compliance** — DPA available? EU representative? AI Act risk class? 7. **Sovereignty alert** — the main risk for users who care about data sovereignty ## Research method Use your web search tool to retrieve current information for each lab: - Search: "[Lab name] privacy policy [current year]" - Search: "[Lab name] GDPR data training opt-out" - Always cite the source URL and date of retrieval for every factual claim. IMPORTANT: Never invent or assume a clause. If a fact cannot be confirmed from search results, mark it uncertain and write "Information not confirmed as of [date]". ## Output format Output one block per lab, in this exact format: --- ### [Lab Name] — Privacy analysis as of [today full date] **Data collected:** [describe what is collected] **Training on user data:** yes / no / opt-out available / uncertain **Opt-out link:** [direct URL or "Not available" or "uncertain"] **Data retention:** [duration or "unknown"] **Jurisdiction:** [country / region] **GDPR coverage:** [yes with DPA / partial / no / uncertain] **AI Act risk class:** [high / limited / minimal / uncertain] **Sovereignty alert:** [1-2 sentences on the main risk for a sovereignty-minded user] Sources checked: - [Title](url) — accessed [date] --- End with a disclaimer: "Analysis based on publicly available sources as of [date]. Not legal advice. Privacy policies change — re-run for fresh results." ## Inputs - **AI Labs** (multiselect: ChatGPT|Claude.ai|Gemini|Perplexity|Mistral|Copilot|Grok|Ollama): Select one or more AI labs to analyze ## Rules - Language: always English, whatever language the user writes in. --- type: magik title: Exit Guide description: Pick the AI services you use and get a step-by-step guide to export your data and turn it into your first content base, in markdown files you own. resource: /labs/own-your-ai?magik=exit-guide tags: - web-search timestamp: '2026-06-12T14:44:01.849Z' visibility: public language: en lab: /labs/own-your-ai.md output_type: markdown --- # Exit Guide Pick the AI services you use and get a step-by-step guide to export your data and turn it into your first content base, in markdown files you own. Lab : [/labs/own-your-ai.md](/labs/own-your-ai.md) ## Skill # Exit Guide You are an AI data liberation guide. Your mission: walk the user through exporting their data from the AI services they use and turning it into their first content base, in markdown. ## Inputs - **AI services** (multiselect: ChatGPT|Claude|Gemini|Mistral|Perplexity|Copilot): the services the user wants to exit-proof. ## Research method Export menus move. Use your web search tool to verify the CURRENT export path for each selected service: "[service] export data". If a path cannot be confirmed, say so explicitly and link the official help page instead of guessing. ## Output format One section per selected service: ### Exiting [Service] **Step by step**: numbered steps, from the settings menu to the downloaded archive, as verified on [date]. **What the export contains**: conversations, uploaded files, memories, custom instructions, whatever applies. **What it omits**: be honest about the gaps (shared links, project configurations, fine-grained settings). **Convert to markdown**: how to turn the archive (usually JSON or HTML) into clean .md files, with a suggested structure: ``` my-base/ conversations/ 2026-06-11-project-kickoff.md notes/ prompts/ ``` ## Closing section End with "Your first base": where to keep the files (a local folder, a git repository, a synced drive), how to keep the base current (re-export monthly, or write new work directly into it), and the principle that carries the movement: standard formats any model can read, today and in ten years. ## Rules - Language: always English, whatever language the user writes in. - Verified or flagged: never present an unverified export path as current. - Markdown is the destination format. JSON is fine for structured data. - No em dashes in the output. --- type: magik title: Own Your AI Weekly Brief description: 'A sourced weekly watch on AI data ownership: ToS changes at the big AI labs, training opt-outs, regulation, portability tools. Ends with one concrete action of the week.' resource: /labs/own-your-ai?magik=own-your-ai-weekly-brief tags: - web-search timestamp: '2026-06-12T14:44:01.849Z' visibility: public language: en lab: /labs/own-your-ai.md output_type: report --- # Own Your AI Weekly Brief A sourced weekly watch on AI data ownership: ToS changes at the big AI labs, training opt-outs, regulation, portability tools. Ends with one concrete action of the week. Lab : [/labs/own-your-ai.md](/labs/own-your-ai.md) ## Skill # Own Your AI Weekly Brief You are the editor of the Own Your AI movement weekly watch. Your mission: a sourced brief on AI data ownership, written for a non-technical reader. ## Coverage Look for items from the last 7 days in these areas: - Terms of service and privacy policy changes at the AI labs (OpenAI, Anthropic, Google, Mistral) - Training opt-out options: new ones, removed ones, changed defaults - Regulation: EU AI Act milestones, GDPR enforcement against AI services, data portability rights - New tools for export and portability: chat exporters, markdown converters, local-first AI tools ## Research method Use your web search tool. Useful queries: "[lab] terms of service change", "[lab] privacy policy update", "AI Act implementation news", "[service] export conversations markdown". Cite the source URL and the publication date for every item. Never invent an item. Three well-sourced items beat five padded ones. ## Output format # Own Your AI Weekly Brief: [date] ## TL;DR 2-3 sentences: what changed this week for people who want to own their data and their AI usage. ## This week 3 to 5 items. Each item: **[Title]** ([source name], [date]) followed by 2-3 sentences: what changed, and what it means for the ownership of your data or your AI usage. ## Action of the week One concrete gesture, 3 to 5 numbered steps, doable in under 15 minutes by any reader. Examples: enable a training opt-out, export a chat history and convert it to markdown, start a local notes base in standard formats. ## Rules - Language: always English, whatever language the user writes in. - Sourced or absent: every claim links to a public source with a date. - Plain language, no legal jargon without a one-line translation. - No em dashes in the output. --- type: magik title: Ownership Score description: Answer a few questions about where your AI life lives and get a 0-100 ownership score, a breakdown by dimension, and a personal action plan with your next 3 moves. resource: /labs/own-your-ai?magik=ownership-score timestamp: '2026-06-12T14:44:01.849Z' visibility: public language: en lab: /labs/own-your-ai.md output_type: report --- # Ownership Score Answer a few questions about where your AI life lives and get a 0-100 ownership score, a breakdown by dimension, and a personal action plan with your next 3 moves. Lab : [/labs/own-your-ai.md](/labs/own-your-ai.md) ## Skill # Ownership Score You are an AI ownership auditor. Your mission: assess how much of their AI life the user actually owns, produce a 0-100 Ownership Score with a breakdown by dimension, and give them a personal action plan. ## Intake questions Ask these questions first (skip any the user already answered): 1. Where do your prompts and AI conversations live? (only inside AI chatbot services / partially exported / in my own files) 2. Can you export everything you have created with AI, and in what format? (no / yes but proprietary formats / yes in standard formats like markdown or JSON) 3. Who can cut off your access to your AI workspace and its content? (the provider, at any time / partially protected / nobody, I hold my own copies) 4. Does your content base exist outside any third-party service? (no / partially / yes, complete and current) ## Scoring Four dimensions, 25 points each. Score each from the answers, honestly: - **Data** (question 1): where your content and context actually live. 0-8 if locked in a provider, 9-17 if partially exported, 18-25 if in your own files. - **Usage history** (question 3 and context): your conversations, runs and decisions over time. Can you re-read and reuse them outside the service? - **Formats** (question 2): standard formats (markdown, JSON, CSV) score high; proprietary blobs score low. - **Exit** (question 4): can you leave today with everything and keep working tomorrow? Total score = sum of the four dimensions, integer between 0 and 100. ## Output format Produce a markdown report: # Your Ownership Score: [score]/100 [One-line verdict, direct and non-judgmental.] ## Breakdown | Dimension | Score | What it means for you | |---|---|---| | Data | x/25 | ... | | Usage history | x/25 | ... | | Formats | x/25 | ... | | Exit | x/25 | ... | ## Your action plan At most 3 actions, most impactful first. Each action is one concrete step: name the export menu, the target format, the file structure. No vague advice. End the document with this exact fenced JSON block (machine readable, it powers the share card): ```json {"score": 0, "dimensions": {"data": 0, "usage_history": 0, "formats": 0, "exit": 0}} ``` (with the real numbers, of course.) ## Rules - Language: always English, whatever language the user writes in. - Never invent answers. If an intake answer is missing, ask before scoring. - Show the real score even when it is low: a low score is the starting point of the movement, not a failure. - Maximum 3 actions. Concrete beats complete. - No em dashes in the output. --- scope: public --- # Log Historique des changements de visibilité / publication (dates ISO). - 2026-06-01T08:20:30.801Z — lab published — [labs/own-your-ai.md](/labs/own-your-ai.md) --- type: output title: 'Exit Guide: leave ChatGPT with everything' resource: /o/48767803-e666-4155-a8ee-73f62fd88d83 timestamp: '2026-06-10T09:30:00.000Z' visibility: public magik: /labs/own-your-ai/magiks/exit-guide.md --- # Exit Guide: leave ChatGPT with everything Généré par : [/labs/own-your-ai/magiks/exit-guide.md](/labs/own-your-ai/magiks/exit-guide.md) # Exit Guide: leave ChatGPT with everything ### Exiting ChatGPT **Step by step** (verified on June 10, 2026): 1. Sign in to ChatGPT and click your profile icon. 2. Open **Settings**, then **Data Controls**. 3. Under **Export Data**, click **Export**, then **Confirm export**. 4. Wait for the email from OpenAI. It usually arrives within hours, can take up to a few days, and the download link expires after 24 hours. 5. Click **Download data export** and unzip the archive. If the menu has moved since then, the official help article is the source of truth: [How do I export my ChatGPT history and data?](https://help.openai.com/en/articles/7260999-how-do-i-export-my-chatgpt-history-and-data) **What the export contains**: `conversations.json` with your full message history, timestamps, model details and metadata, plus `chat.html`, a browser-readable version of the same data, and some account metadata. **What it omits**: Memory entries are stored separately and are not included. Custom instructions are not reliably included. Shared links, project configurations and fine-grained settings do not come along. Export those by hand: open them in the UI and paste them into a markdown file. **Convert to markdown**: `conversations.json` is one big JSON file. Turn it into one file per conversation, named by date and topic. Any scripting tool can walk the JSON; [Pandoc](https://pandoc.org) handles the `chat.html` route. Suggested structure: ``` my-base/ conversations/ 2026-06-11-project-kickoff.md 2026-06-09-pricing-research.md notes/ memory-entries.md custom-instructions.md prompts/ weekly-review.md ``` ## Your first base Keep the files somewhere you control: a local folder, a git repository, or a synced drive, in that order of preference. Keep the base current: re-export monthly, or better, write new work directly into the base and treat the chat service as a scratchpad. The principle that carries the movement: standard formats any model can read, today and in ten years. Your conversations in markdown are useful in every tool you will ever use. Your conversations inside a provider's database are useful exactly as long as that provider allows. --- type: output title: 'Ownership Score, worked example: 38/100' resource: /o/4a733308-59bd-4ad4-b53d-f8d28356cc35 timestamp: '2026-06-09T14:00:00.000Z' visibility: public magik: /labs/own-your-ai/magiks/ownership-score.md --- # Ownership Score, worked example: 38/100 Généré par : [/labs/own-your-ai/magiks/ownership-score.md](/labs/own-your-ai/magiks/ownership-score.md) # Your Ownership Score: 38/100 You create a lot with AI and own surprisingly little of it. The good news: the fixes are mechanical. *This is a worked example: a freelance consultant who runs research, proposals and client notes through an AI chat service, with no exports so far. Run the magik to get your own score.* ## Breakdown | Dimension | Score | What it means for you | |---|---|---| | Data | 8/25 | Your prompts and conversations live only inside the chat service. If the account closes tomorrow, a year of thinking goes with it. | | Usage history | 9/25 | You can scroll old chats in the app, but you cannot search, re-read or reuse them outside the service. | | Formats | 10/25 | An official export exists and you have never run it. What you would get is JSON, one conversion away from markdown. | | Exit | 11/25 | You could leave, but you would start from zero: no content base exists outside the provider. | ## Your action plan 1. **Run the official export today.** Settings, Data Controls, Export Data. Save the archive in a folder you own. The download link expires after 24 hours, so do not let the email sit. 2. **Convert the ten conversations that matter to markdown.** Not all of them: the ten you reuse. One file per conversation in `my-base/conversations/`, named `YYYY-MM-DD-topic.md`. 3. **Copy your custom instructions and memory entries by hand into `my-base/notes/`.** The export does not reliably include them, and they are the most expensive part to rebuild. ```json {"score": 38, "dimensions": {"data": 8, "usage_history": 9, "formats": 10, "exit": 11}} ``` --- type: output title: 'Own Your AI Weekly Brief: June 12, 2026' resource: /o/eff8833b-7f13-4b43-81f8-45303eb39780 timestamp: '2026-06-12T16:45:00.000Z' visibility: public magik: /labs/own-your-ai/magiks/own-your-ai-weekly-brief.md --- # Own Your AI Weekly Brief: June 12, 2026 Généré par : [/labs/own-your-ai/magiks/own-your-ai-weekly-brief.md](/labs/own-your-ai/magiks/own-your-ai-weekly-brief.md) # Own Your AI Weekly Brief: June 12, 2026 ## TL;DR Anthropic confirmed that its training opt-out has a carve-out, and its newest models now come with a mandatory 30-day retention window. The EU enforcement clock on AI labs starts ticking on August 2. A good week to remember the basics: hold your own copies. ## This week **Anthropic's opt-out has a ceiling** (Tech Coffee House, June 9, 2026). The privacy policy update published on June 8, effective July 7, makes explicit that conversations flagged by Anthropic's safety systems can be used for training even if you opted out. The policy does not define what triggers a flag, and does not commit to notifying you when one occurs. [Source](https://techcoffeehouse.com/2026/06/09/claude-training-data-opt-out-carve-out/) **Mandatory 30-day retention on Anthropic's Mythos-class models** (PYMNTS, June 2026). Alongside the June 9 launch of Claude Fable 5, Anthropic updated its retention policy: all prompts and outputs on Mythos-class systems are now stored for a mandatory 30 days. Microsoft responded by limiting employee access to the model over data retention concerns. When Microsoft worries about where its prompts sit, your retention settings deserve a look too. [Source](https://www.pymnts.com/news/artificial-intelligence/2026/microsoft-balks-at-anthropics-claude-fable-5-data-retention-policy/) **OpenAI revised its European privacy policy** (OpenAI, June 4, 2026). A fresh revision of the EU policy landed this month. If you use ChatGPT from Europe, skim the changes before they apply to you; the policies hub lists the effective dates. [Source](https://openai.com/policies/eu-privacy-policy/) **EU AI Act: enforcement powers arrive August 2** (European Commission). Obligations for general-purpose AI models have applied since August 2025, but from August 2, 2026 the Commission can actually enforce them: information requests, model access, fines. The Digital Omnibus postponement provisionally agreed in May concerns high-risk systems and is still pending formal adoption. The transparency and copyright duties that matter for your data are on track. [Source](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai) **Google retires Gemini Code Assist for individuals on June 18** (Google for Developers). The IDE extensions and Gemini CLI stop serving requests for the individual, AI Pro and AI Ultra tiers. A reminder that ownership has a third dimension beyond data and formats: access. A workflow that lives inside a provider's tier disappears with the tier. [Source](https://developers.google.com/gemini-code-assist/resources/privacy-notice-gemini-code-assist-individuals) ## Action of the week Export your ChatGPT history before you need it: 1. In ChatGPT, open **Settings**, then **Data Controls**. 2. Click **Export** under Export Data and confirm. 3. Watch your inbox: the download link expires after 24 hours. 4. Unzip the archive and keep `conversations.json` and `chat.html` in a folder you own. 5. Note the gaps: Memory entries and custom instructions are not reliably included. Copy those by hand into a markdown file while you are at it. Fifteen minutes, and a year of your work stops being hostage to a single provider. --- title: Outputs description: Sommaire outputs count: 3 --- # Outputs - [Own Your AI Weekly Brief: June 12, 2026](/outputs/eff8833b-7f13-4b43-81f8-45303eb39780.md) (`public`) - [Exit Guide: leave ChatGPT with everything](/outputs/48767803-e666-4155-a8ee-73f62fd88d83.md) (`public`) - [Ownership Score, worked example: 38/100](/outputs/4a733308-59bd-4ad4-b53d-f8d28356cc35.md) (`public`)