Ask yourself: “What am I sharing?”

Once it’s sent, it’s out of your hands

Be careful about sharing personal or sensitive data with AI chatbots. Things you share with them may not stay private.

Most people prefer to keep some things private. These might be things which are personal and sensitive. They might also be things that could be used to identify you, such as your full name, home address, date of birth, or identity numbers. If this kind of information gets into the wrong hands it could be used to steal your identity.

Before you share personal details with an AI chatbot, ask yourself "would I be happy for a stranger to see this?”

AI chatbots work by sending your message through the internet to the company’s computers. These computers process your message, and then send a response back to you.

The messages and other information you send to an AI chatbot do not disappear when the conversation ends. This is because the company stores it on their computers, often for a long time.

This matters for two reasons:

  1. Stored data can leak out. Any company which holds a large amount of personal data can be hacked or it might release information by mistake.
  2. Other people may see it without you wanting them to. If someone else can get into your account, they can read your past conversations.

Many AI companies use the information you provide to improve their chatbots. Some companies let you turn this off, or say when you would rather keep things private.

You can find out more about how chatbot companies use your information by reading their privacy policies. Below are links to the privacy policies for some of the most widely used AI chatbots (links last checked 17th September 2026). Remember – these policies might change over time.

There are two things you need to consider:

  1. Most AI chatbots won't have access to your medical history or information about your lifestyle. This means that when you ask them questions, their responses might be generic and less suited to you. If you share information about your health with them, you might get more tailored responses.
  2. On the other hand, sharing lots of information with a chatbot could put your privacy at risk.

It's your choice how much information you share. There is no right answer. It's up to you how you balance these issues.

Many chatbots have memory. This means that the chatbot will store information about you from past messages you have sent and use this in future replies. This is not the same thing as human memory.

Here is an example of how this might work: if you have told the chatbot about your job in the past, it may know that you work in an office. A few weeks later, if you then ask about back pain, it might give advice about how to improve your posture when sitting at a desk.

Over time, the chatbot may have stored quite a lot of information about you. Some people find this useful, but others might prefer to turn this off in the settings.

Most chatbots will store some of the information that you share with them. This might be as memory (described above). It might also be in the chat history, which shows all the previous conversations you've had with the chatbot.

Some AI chatbots have a setting you can turn on so that information you share in a particular conversation does not appear in your chat history. Different chatbots have different names for this setting, and each chatbot does it slightly differently. ChatGPT, Google Gemini and Microsoft Copilot call this ‘temporary chat’. Claude calls this ‘incognito chat’. You can learn more about these by visiting the chatbot companies’ websites.

Even if you use temporary / incognito mode, most chatbots will still store information and questions you send for a short period. This might be for legal and safety reasons.

If you share a chatbot account with other people, they might be able to see the messages you have sent to the chatbot in the past. This might include personal or sensitive information you would rather keep private.

Many AI chatbot companies have policies which say you should not share your account with other people.

It’s very worrying when someone close to you is unwell. AI chatbots might help you understand what is going on. Microsoft researchers found that about 1 in 7 personal health questions sent to Copilot were about other people rather than the user.1

Asking general questions is usually fine. But think before sharing anything specific or personal about someone else. Do you need their permission first? In particular, always ask before sharing something they told you in confidence, including test results.

Sometimes the person you are worried about can’t make this decision themselves. This might be because they are seriously unwell, or because they are a very young child. In these cases, think carefully about whether sharing their information with an AI chatbot is in their best interests. Many older children and people with a learning disability can decide for themselves whether they are happy for their information to be shared. Where possible, ask them directly.

Two things can protect your account from being hacked into. Use a strong password, and if the chatbot offers it turn on multi-factor authentication (MFA). MFA means that even if someone steals your password, they still can't get into your account without a security code sent to your phone or other smart device.

Think carefully before using your work’s AI chatbot for personal or health questions. Depending on how it is set up, your employer may be able to see anything you type. If you want to ask about your own health, it is safer to use a personal account on your own device.

You should also be careful about using your own personal chatbot for your work. Many employers will have specific policies about the use of AI chatbots. Some employers may provide access to a specific chatbot that is approved for use.