AI privacy and data safety basics
Every prompt you send to an AI tool is data leaving your device. A few simple habits — checking terms, minimising what you share, and using the right plan for sensitive work — keep that risk low.
What happens to your prompts
When you use a cloud-based AI tool, your prompt is sent to the provider's servers, processed there, and often stored for some period, and on some plans it may also be used to improve the provider's models. That means anything you paste — a document, a customer email, a photo, or a piece of code — could be retained beyond your conversation, depending on the provider and the plan. Free consumer plans and paid business plans often differ in exactly this respect, with business plans more commonly offering stronger commitments not to train on your data. Before using an AI tool for work, read the provider's privacy terms for the plan you will actually use, and check whether your employer or client has rules about which tools are approved. If you cannot find a clear answer, treat the tool as unsuitable for sensitive material.
What never to paste into a general AI tool
Some information should never go into a general chatbot or AI tool: passwords, API keys, and security credentials; full customer records, addresses, or payment details; medical or financial information about identifiable people; and confidential business material such as unpublished prices, contracts, or trade secrets. Even if a tool promises privacy, a pasted secret can be copied into a conversation history, a shared screen, or an exported file, and removing it afterwards is difficult. If you need help with a document that contains sensitive details, remove or replace those details with placeholders before pasting, and reinsert them afterwards in your own files. This habit takes seconds and prevents the most common and most serious AI privacy mistakes.
Accounts, settings, and history
Check the privacy settings of the AI tools you use: many let you turn off training on your conversations, delete chat history, or set how long data is retained, but those settings are not always on by default. Review them when you first sign up and again when a tool updates its terms, since defaults can change. Use a strong, unique password and two-factor authentication on AI accounts that hold your conversations, since a compromised account exposes everything you have discussed. If you share a device or an account with colleagues or family, remember that conversation history may be visible to them unless you use separate accounts or delete sensitive chats. Small settings checks, done once, protect a great deal of personal and business information.
Photos, voice, and biometric data
Photos, voice recordings, and other biometric material deserve extra caution, because unlike a password, you cannot change your face or voice if they are misused. Avoid uploading photos of other people, especially children, into AI tools without a clear reason and permission, and think twice before using face or voice features that create a lasting digital copy of you. Voice cloning and face tools can be used legitimately with consent, but a recording uploaded casually for a fun effect may be retained, analysed, or used in ways you did not expect. Check what a tool does with uploads, how long it keeps them, and whether you can delete them, and prefer tools that process sensitive media locally or offer clear deletion. When in doubt, do not upload it.
AI at work and in school
Workplace and school rules about AI often exist for privacy reasons, not just for academic honesty: an employer may forbid pasting client data into external tools, and a school may forbid uploading other students' work or unpublished research. Check your organisation's policy before using AI for work or study, and ask if the policy is unclear, since guessing wrong can have professional consequences. If your employer provides an approved AI tool, use that one for work material rather than a personal account, since the approved tool usually comes with stronger privacy terms and central controls. Keep work and personal AI use in separate accounts where possible, so a personal conversation never sits alongside confidential work material in the same history.
Children and vulnerable people
Extra care is needed when AI tools are used by or about children and vulnerable people, since their data deserves stronger protection and they may not understand what sharing means. Check the age limits and parental controls of any AI tool a child will use, supervise its use, and avoid uploading children's photos, school records, or personal details into general AI tools. Be cautious with AI companions and chatbots aimed at children, and talk to children about not sharing their name, school, address, or photos with a chatbot. If you care for someone vulnerable, apply the same rule: do not share their personal, medical, or financial details with an AI tool unless there is a clear, necessary reason and an appropriate, approved tool for the job. Privacy here is a duty of care, not just a setting.
A simple data-safety routine
A routine that keeps AI use safe without slowing you down looks like this: before pasting, ask whether the material contains secrets, personal data, or confidential business information, and remove anything that does; check the tool's privacy terms and settings once, and use a business or approved plan for work material; delete sensitive conversations when you no longer need them; and keep strong account security on every AI service you use. Teach the same routine to anyone who uses AI tools in your business or household, since one careless paste can undo everyone else's caution. AI tools are genuinely useful, and they do not have to be risky — but the safety comes from your habits, not from assuming the tool will protect you by default.