← All AI guidesMake sense of AI

Staying safe when using AI

Prefer listening? Have this guide read aloud.

AI can be useful, interesting and even fun. A few simple habits can help you use it safely and confidently.

You don't need to be an expert or worry about every question you ask. But just as you would take care when shopping online or sharing information on a website, it's sensible to take a little care when using AI.

1. Think before sharing personal information

Chatting with an AI assistant can feel like a private conversation, but you're using an online service. What you type may be stored or processed by the company providing it.

Different services have different privacy policies and settings. As a general rule, avoid sharing:

  • Passwords, PINs or security codes.
  • Bank account or credit card details.
  • Passport, driver's licence or identity document numbers.
  • Private medical information or other sensitive personal details.
  • Personal information about family, friends or others without their permission.

Instead of entering your bank statement, you could ask:

Can you explain what the different sections of a bank statement mean?

Ask yourself: Do I really need to share this information to get the help I want?

2. Understand that AI conversations aren't always private

Some AI services save conversations so you can return to them later. Depending on the service and your settings, conversations may also be used to improve the AI.

Look for Settings, Privacy or Data Controls to learn what choices you have. Deleting a conversation from your screen doesn't necessarily mean every copy is immediately deleted.

You don't need to stop using AI because of this. Just be thoughtful about what you share.

3. Be aware of AI-generated scams

AI can create realistic-looking pictures, videos, messages and even voices. Most people use these for ordinary or creative purposes, but scammers can use them too.

  • A convincing email pretending to be your bank.
  • A fake photograph or video of a well-known person promoting an investment.
  • A computer-generated voice that sounds like someone you know.
  • A realistic-looking website pretending to be a trusted organisation.

A convincing voice, photograph or video is no longer proof that something is genuine.

What if someone sounds exactly like a family member?

Imagine a call from someone who sounds like your son, daughter or grandchild. They say they're in trouble and urgently need money.

Even if the voice sounds familiar, don't feel pressured to act immediately. Hang up and contact your family member using a number you already know, or check with another trusted family member.

Taking time to check is sensible, not rude. Read our Someone called me guide for more help.

If you've already sent money or shared sensitive information, act promptly by contacting your bank or the relevant service.

4. Be careful with links and recommendations

An AI assistant might suggest a website, product, service or app. That doesn't automatically mean it's safe, genuine or suitable for you.

Before downloading something, making a payment or entering personal details, check the genuine website, look for independent information and ask someone you trust if you're unsure.

5. Remember that AI isn't a person

AI assistants can be friendly, encouraging and easy to talk to. Some can remember details from earlier conversations. This can feel surprisingly personal.

But an AI assistant isn't a human friend, doctor, financial adviser or other qualified professional. It doesn't know everything about your circumstances and can make mistakes.

It's fine to enjoy chatting with AI. For important personal matters, it's worth talking with people you trust or someone appropriately qualified.

AI can be a helpful tool, but it shouldn't replace your own judgement or the people who care about you.

6. You are always in control

You don't have to answer every question an AI assistant asks, follow every suggestion or keep using a service that makes you uncomfortable.

If an answer seems strange or misleading, question it, check elsewhere or stop. Remember that AI can make mistakes when checking its own answers.

What if something has already gone wrong?

If you've shared a password, sent money to a scammer or given someone access to your device, don't feel embarrassed. The important thing is to take appropriate action promptly.

Visit Soilleir's Stay Safe Online guidance for practical next steps. You can also ask a trusted friend or family member for help.

One last thing

You don't need to be frightened of AI to use it safely. Be curious, ask questions and explore things that interest you. Keep sensitive information private, check things that matter and take your time when something doesn't feel right.

AI should help you feel more capable and confident, not less.

Sources and further reading

These links offer more information if you would like to explore the topic.

Was this guide helpful?

We welcome suggestions, corrections and ideas for making this guide clearer.

Send feedback about this guide