All guides
For everyday AI users checking default settings, data sharing and privacy choices.
- A cautious starting point for AI privacy settingsA cautious starting point is to use temporary chat so a conversation stays out of history while temporary, while separately checking that turning off model improvement does not delete or hide saved chats.
- A data-control review for signed-in and signed-out AI toolsA practical review of data controls for signed-in and signed-out AI tools, centered on the fact that available controls depend on sign-in status, plan, and workspace settings.
- A privacy baseline for family, school, and work useA privacy baseline for family, school, and work use starts with the fact that a managed-account administrator may access and control associated data and that available controls depend on whether the person is signed in, the plan, and workspace settings.
- Backups and deletion: what to ask before assuming removalArchiving keeps a chat in the account under saved-chat retention rules, while deleting a chat does not remove files saved separately in Library.
- Before an AI tool joins a workflow, decide its data boundaryThis guide explains how to check app permissions separately from general conversation controls and how connected-app access can be removed before an AI tool enters a workflow.
- Before granting connected-app access, map the data pathThis guide separates app-specific permissions from general conversation settings and identifies the data-handling questions to confirm before access is granted.
- Can opting out of training still leave conversational logs?Turning off model improvement does not delete or hide saved chats, so chat history and applicable retention settings must be checked separately.
- Default data sharing for a new AI account: questions before opt-inThis guide explains what can and cannot be verified about default data sharing for a new account, including that eligible accounts can request a data export and app permissions are separate from general conversation settings.
- Default settings for AI use in health, financial, and legal contextsThis guide explains what the available guidance establishes about AI defaults: files may be stored separately, while content used in a conversation may remain accessible for that conversation’s life.
- File uploads: checking retention, sharing, and deletion choicesThis guide explains how to check file retention, sharing, and deletion controls, including why deleting a chat may not remove a separately saved file.
- Helpful personalization versus tighter data controlExplains how Memory can use past chats for personalization and why turning off model improvement does not delete or hide saved chats.
- How to tell whether a setting is preselected, unavailable, or administrator-managedA setting can already be selected, be unavailable, or be controlled by workspace settings, so its displayed state and account context must be checked before assigning a cause.
- Keeping data for service operation versus reusing it for model developmentExplains that turning off model improvement does not delete saved chats, while a temporary-chat copy may be retained for up to 30 days for safety.
- Memory across conversations: a default to consider, not a default to acceptThis guide explains that Memory can use relevant details from past chats when enabled, but turning it off does not delete those chats.
- Personalization and model improvement: why separate settings matterMemory can use chats and other available sources for personalization, and app permissions are separate from general conversation settings; neither cited statement establishes model improvement.
- Private links or public links? Choosing a defaultA practical guide to choosing a private or public link default, with the key fact that anyone with a link created from a personal account can view the shared conversation or scheduled task.
- Sensitive prompts: when unclear retention settings are a reason to stopUnclear retention settings can justify stopping before a sensitive prompt is entered because available controls vary by account context, while disabling model improvement does not delete saved chats.
- Shared links: who can open a saved AI conversation?A shared personal-account link can be opened by anyone who has it, while archiving keeps the chat under saved-chat retention rules.
- Should conversation history be on by default?The documentation does not establish a universal default, but it says regular chats remain in history until deletion or workspace retention removes them.
- Terms, settings, and consent: where the choices should be checkedA safe-defaults checklist for checking model improvement, saved chats, and connected-app access without treating them as the same choice.
- Testing whether an AI privacy setting actually took effectA valid test matches the control’s documented effect, because turning off model improvement does not delete or hide saved chats.
- The file-upload defaults worth reviewingA review of separate file retention and chat deletion, including the key fact that deleting a chat does not delete files separately saved in Library.
- Using AI on a shared device without carrying over old accessExplains how to review active sessions and trusted devices in security settings and log out of individual or all sessions when using AI on a shared device.
- What a private label does not settle on its ownA private label alone does not establish who can open a shared link or how app permissions differ from controls applying to all conversations.
- What to check when a privacy control applies only to new conversationsWhat to verify when a privacy control applies only to new conversations, including that turning off model improvement does not delete or hide saved chats.
- When a cautious default calls for human reviewA cautious default calls for human review before a personal-account link is shared, because anyone with that link can view the shared conversation or scheduled task.
- When an administrator manages the privacy controlsAdministrator-managed privacy settings can govern retention, sharing, data controls, and access to organization-managed services and features.
- When personalization deserves a second lookEnabled Memory can draw on past chats and other available sources, but turning it off does not delete past chats.
- Which AI settings deserve review before using a shared device?Explains which managed AI account controls to review on a shared device, including administrator access to associated data and authority over retention, sharing, and available features.
- Account deletion: questions to ask before clicking deleteA pre-deletion checklist covering the permanence of account deletion and the saved-chat retention rules that apply when a chat is archived.
- Account recovery details to review before switching devicesThe account email or phone number used to start password reset and the reset email instructions are confirmed details to review before switching devices.
- Activity history: review it, clear it, or keep it?Explains when to review, clear, or keep activity history, and notes that available data controls depend on sign-in status, plan, and workspace settings.
- Audio and video inputs: which privacy settings apply?Explains why archiving does not delete associated audio or video clips and identifies the general data controls readers should review.
- Can an exported account file reveal more than expected?Explains that eligible signed-in users can request a data copy, while the available guidance does not confirm every exported item and notes that audio and video sharing may have separate controls.
- Can I get a copy of my AI conversation data?Eligible account users can request a data copy through settings, but separately stored files may require an additional scope check.
- Connected apps: checking which services can reach your accountExplains that connected-app permission settings determine when approval is requested and what readers must confirm about current access and removal options.
- Conversation history: save it, pause it, or delete it?A decision guide to saving, pausing, or deleting conversation history, including the key fact that disabling model improvement does not keep new conversations out of chat history.
- Finding download, deletion, and account closure controlsExplains that eligible signed-in users can request a data copy from settings while distinguishing confirmed deletion controls from unverified account-closure details.
- Hidden, disabled, or administrator-controlled: what a missing setting may meanA missing data control does not establish why it is unavailable: the cited documentation covers model training, chat history, exports, and account deletion but does not diagnose absence.
- How to check who can see shared AI chat linksReview the audience before sharing because anyone permitted to open a shared link can view its content.
- How to find the control for response personalizationOpenAI places ChatGPT memory controls under Settings > Personalization, while availability can vary by plan, region, platform, and workspace settings.
- How to inspect the data an AI service says it collectsThis guide explains how eligible users can request a data copy from account settings and why managed-account users must also confirm administrator control.
- Importing old chats: what to check before bringing them inAn exported conversation file can be uploaded to a new conversation in another personal account for reference, but it does not recreate old chats in the destination account's history.
- Login alerts versus marketing emails: which controls to separateA guide to separating login alerts from marketing-email preferences, including alerts for unrecognized locations or devices and verification by app approval or emailed code.
- Private chat labels: what they do and do not tell youA Temporary Chat stays out of chat history while temporary, but OpenAI says a copy may be kept for up to 30 days for safety purposes.
- Public profile or private profile? Read the AI setting firstA personal profile is private by default, while a public profile can be viewed by anyone signed in to the service.
- Temporary chat, saved history, and deleted history: what to verifyA practical checklist for temporary, saved, and deleted history, including that a temporary chat may still be retained for safety for up to 30 days.
- The AI account settings that deserve a privacy checkA checklist for the two ChatGPT settings areas OpenAI identifies for privacy review: model-improvement preferences and memory controls.
- The difference between account deletion and deleting individual chatsAccount deletion is permanent, while an individual conversation can be removed without deleting the account.
- Training opt-out, retention, and human review: where the controls differModel-improvement opt-out can prevent new conversations from training models while leaving them in chat history, but human-review controls remain unspecified.
- What gets kept after an AI conversation ends?Regular chats remain in history until deletion or workspace retention removes them, while archived chats follow saved-chat retention rules.
- What to verify before enabling cross-conversation memoryA practical checklist for confirming sign-in status, plan, workspace settings, and possible separate audio or video controls before enabling cross-conversation memory.
- Who can access data submitted to an AI service?Data associated with a managed account may be accessible to and controllable by the account administrator.
- Work-managed versus personal AI accounts: what to verifyA practical comparison of work-managed and personal AI accounts, including why disabling model improvement does not also remove new conversations from chat history.
No articles have been published yet. The section structure is ready.