Jump to a section
ChatGPT is useful for everyday drafts and ideas when you choose what to share and check important answers. Leave confidential information out of personal chats, review Data controls and treat Temporary Chat as a limited privacy choice rather than a confidential service.
The short answer
ChatGPT can be useful for ideas, explanations and first drafts. Use it for jobs where you can judge the result, choose what you share and check information that matters.
It is not a private adviser or a guarantee that an answer is correct. A cautious starting point does not need to be complicated: use made-up details and one ordinary question.
How to ask an AI chatbot a good question (the four-part trick)
What happens to what you type
On a personal ChatGPT account, the Improve the model for everyone setting controls whether new conversations may be used to improve OpenAI's models. Turning it off does not remove chats from your history.
Temporary chats are not used for model improvement while they remain temporary and do not create or update memories. OpenAI may retain a copy for up to 30 days for safety. If you save one as a regular chat, your normal settings apply.
- Open your account menu and choose Settings, then Data controls.
- Review Improve the model for everyone and turn it off if you do not want new chats used for training.
- For an unsaved conversation, start a Temporary chat. Review its personalisation choice before you send anything.
Is ChatGPT safe for confidential information?
For an ordinary personal account, leave confidential information out. That includes bank and card details, passwords, security codes, ID documents, private legal material, medical records and other people's personal information.
Our practical recommendation is to give the shape of the problem instead. Ask about 'a letter asking for £[AMOUNT] by [DATE]' and remove names, addresses, account numbers and references. When the exact figures matter, add them yourself after the draft is written.
Turning training off changes one use of your data; it does not make a chat a confidential advice service. OpenAI also says feedback such as a thumbs-down can allow the associated conversation to be used for model improvement. Read that choice before sending feedback about sensitive material.
What never to tell an AI chatbot (and what's fine)
AI chatbots and your privacy: what they keep, who sees it, and how to take control
Can you trust its answers?
OpenAI says ChatGPT can give confident but incorrect answers and invent references. A cited source is a place to check, rather than proof that the answer has already been verified.
Open the page, find the sentence or figure and check the date. If a response affects a deadline, money, health or legal rights, confirm it with the relevant organisation or qualified person.
AI chatbots make things up. Here's how to check what they tell you
What about teenagers?
OpenAI offers linked parental controls for supported teen accounts. These can manage selected features and quiet hours. They do not give a parent access to the teenager's conversations or real-time activity.
Set up the account together and discuss what your teenager should do when an answer is upsetting or seems wrong. Treat the controls as one part of that arrangement, rather than a substitute for the conversation.
Use the real service
Open the service directly instead of following an urgent payment message or an advert that claims to be support. If you install an app, check the publisher carefully before signing in.
You do not need to send somebody your password, one-time sign-in code or card details to ask a chatbot a question. If you have already shared a password, change it with the real service.
AI scams explained: what's changed, and the habits that still beat them
A useful first try
Try a low-stakes task with no personal information: 'Rewrite this invitation in a warmer tone, under 80 words: [YOUR DRAFT].' Read the result and edit what does not sound like you.
That is enough to find out whether it helps you. You can decide later whether to use files, memory, voice or a paid plan.
Choose your next step
Pick the guide for the tool or task you use.
Is Gemini safe? Privacy checks for a personal Google account
Five things that make ChatGPT safe to use
- Only use chatgpt.com or the official ChatGPT app
- Decide on 'Improve model for everyone' in Settings, then Data controls
- Use Temporary Chat for anything personal
- Keep passwords, account numbers and ID details out of every chat
- Check anything important, especially health, money and legal answers
Sources
These sources support the guide. Tool settings, prices and policies change, so dated source boxes show when those details were checked.
- OpenAI: managing parental controls in ChatGPT
- OpenAI: does ChatGPT tell the truth?
- OpenAI: data controls in ChatGPT
- OpenAI: temporary chat in ChatGPT
AI helps with research and drafting. Dan checks the sources and is responsible for what is published.
How we check