Everyday AI users · 19 pages · Digital PDF
Know what your AI assistant remembers — and what you control.
Understand training, memory, connected apps and the information you should not upload.
What’s included
- ChatGPT memory and training options
- Gemini activity
- Claude data controls
- Copilot privacy
- Connected apps, MFA and redaction
- AI privacy audit and printable checklist
Understand what you control
Protect Your Personal Data in AI concentrates on information handling and account review. It is intended for everyday users who want to understand training preferences, history, memory and connections without treating them as interchangeable. A control that changes model-improvement use does not necessarily change visible history or remove a remembered fact. An app connection can bring information into the task from somewhere else. Start by identifying the account type and current provider terms. Consumer, work and school arrangements may differ. The guide is a printable educational aid for asking the right questions, not a guarantee that all conversations are private or a permanent map of changing menus. The related articles link to official documentation so that you can check current labels and exceptions before relying on a setting.
ChatGPT, Gemini, Claude and Copilot
The intended provider coverage includes ChatGPT memory and training options, Gemini activity, Claude data controls and Copilot privacy. Each service has its own terminology and feature availability. Review the controls relevant to your actual account, not an old screenshot or a statement about another plan. Temporary or incognito modes can limit specific uses while still having safety-related handling described in provider guidance. Activity deletion also needs a separate review from future settings. The guide does not declare one assistant completely safe or endorse a particular provider. It aims to make your audit more deliberate: establish what is saved, what can be remembered, what can be accessed and what you can change. Keep the official documentation alongside a printed checklist when a feature has been updated.
Redaction and what not to upload
Many useful tasks can be completed with fictional examples or a short non-identifying extract. Avoid sharing credentials, full identity documents, unnecessary financial details and private records about another person. A spreadsheet or screenshot may expose more than the visible question, including extra sheets, account names and notifications. Review the complete item rather than assuming the assistant ignores everything outside your intended section. Replacing a name is not always anonymisation; a combination of unusual dates, locations and circumstances can still identify someone. The guide’s intended redaction and upload checks help you pause before submission. They do not authorise a sensitive upload at work or guarantee that edited material can no longer identify a person. When the real information is essential, use an approved process suited to that data.
Connected apps and account security
An assistant with email or drive access can reach considerably more context than a standalone prompt. Check the account linked, permission scope and purpose of each connection. Disable ones no longer needed and review permissions in the original service as appropriate. Do not assume disconnecting reverses earlier transfers. Login security remains important too: use unique credentials, supported multi-factor authentication and a protected device. Be cautious of unexpected messages offering a privacy fix or account warning. Navigate to the service independently instead of entering details through an unfamiliar link. The guide’s security coverage supports a practical routine, but no list of settings removes every threat. If a credential was exposed, revoke or change it promptly rather than relying on deletion of a visible conversation.
A privacy audit you can repeat
A useful audit records the account, relevant settings, connections and questions still unresolved. Keep it short enough to revisit after a new app, plan or feature is introduced. Avoid storing screenshots that themselves reveal private information. The intended printable checklist is there to support repetition, not to certify compliance or total safety. Choose the Playbook if you want broader personal scam and deepfake awareness; choose the Parents guide for family discussions and a family agreement; choose Small Business for staff and supplier procedures. The free checklist offers a smaller starting point without mandatory registration. The complete library can suit people with several roles, but you should compare the scope of each guide before purchasing. Better privacy comes from consistently sharing only what a task needs as well as reviewing the software controls.
Frequently asked questions
Is this an official provider guide?
No. BMR Digital publishes independent educational material without affiliation, certification or endorsement from OpenAI, Google, Microsoft or Anthropic.
Will the guide make every AI conversation safe?
No. Settings and terms vary, mistakes and scams remain possible, and no guide guarantees complete safety. Use current official sources and appropriate professional help.
Do I receive a printed book?
No. These are digital PDF downloads. You can save and print the original uploaded documents for your own use.
Can I download without creating an account?
The free checklist needs no account or mandatory email. Verified purchasers use a private order access link for repeat downloads, without a full customer account.
What if my file does not arrive or is faulty?
Use the private order support route or the existing contact form with your order reference. Statutory consumer rights for faulty or misdescribed content are not removed by downloading.
Are changing settings a legal compliance guarantee?
No. Business templates must be adapted, and current data-protection guidance and qualified advice may be needed for your actual use.
Digital PDF download — no physical product
This purchase is for the stated digital PDF content, not a printed book, subscription, consultation or ongoing security service. A genuine preview is taken from the first page of the uploaded guide rather than an invented cover. The page count is checked against the supplied product specification before availability is enabled. You can save the original file and print it for your own use. Device PDF viewers vary, so ensure that your phone or computer can open a standard PDF before relying on it during an urgent task. If a file is missing or does not open properly, contact BMR Digital through the existing enquiry form or the private order support route. Include an order reference where you have one, but never send passwords or full card details.
Payment, immediate supply and your rights
Paid access is granted only after a completed PayPal capture has been checked for the correct order, receiving merchant, GBP currency and server-calculated amount. Failed, cancelled, pending or unverified payments do not provide files. Before payment, immediate supply requires your express consent and acknowledgement that the 14-day cancellation right is lost once supply begins under the applicable digital-content rules. That acknowledgement does not remove statutory remedies for faulty or misdescribed content. BMR Digital does not automatically deny a lawful refund merely because a file has been downloaded. Review the checkout confirmation and contact the business if the content or delivery does not match what you purchased. Live payments remain unavailable until the owner has completed testing and explicitly authorised launch.
Independent guidance and changing products
BMR Digital publishes independent educational material and is not affiliated with or endorsed by OpenAI, Google, Microsoft or Anthropic. Provider names identify the services discussed, not a certification or partnership. Features, account plans and privacy wording can change. Use the official source links in the resource centre to confirm the current position for your account. No guide guarantees complete safety, prevents every scam or automatically makes a UK business compliant. Do not substitute a generated answer or a printable template for qualified medical, legal, financial, security or safeguarding help where needed. The useful outcome is a clearer routine: share less unnecessary information, review permissions, verify unusual requests and keep a responsible person involved in decisions that affect real people.
Need family, privacy and workplace guidance?
Compare the Complete AI Safety Library and its current genuine saving. The bundle includes all four paid PDFs, with each guide downloaded separately after verified payment.
View the librarySupport and delivery questions: contact BMR Digital. Read our existing terms and privacy policy. Paid checkout remains locked until seller details and testing are ready.
Useful AI safety reading
All resourcesChatGPT Privacy Settings: What Each Control Does
Training, history, memory and connected apps are separate ChatGPT controls. Review each one before sharing personal or customer information.
AI Safety for Children: A Parent's Starting Guide
Begin with age rules, private information, supervised chatbot use and calm family conversations about deepfakes and misleading answers.
AI Policy for Small Business: Simple UK Starter Guide
Set approved tools, prohibited data, human checks and incident reporting before staff use generative AI for customer work.
AI Voice-Cloning Scams: How to Check a Fake Call
Do not rely on a familiar voice alone. End the call and independently contact the person or business through a known route.
What Not to Share With AI: Practical Examples
Keep passwords, identity documents and unnecessary personal or confidential details out of AI prompts. Use fictional examples and minimise data.
Gemini Activity & Privacy Settings Explained
Review Gemini Keep Activity, deletion, connected services and account type separately; switching off activity does not mean no processing or retention.
Claude & Copilot Privacy Controls Compared
Compare account type, training preferences, history, memory and connections rather than assuming Claude or Copilot is completely private.