🔒 Stay Safe with AI
AI is a helpful tool, but like any tool, it pays to use it wisely. This page covers everything you need to know to stay safe.
🚫 What Not to Share with AI
AI assistants are not private diaries. Treat them like a helpful stranger. Be friendly, but keep sensitive things to yourself.
Never type these into an AI chat:
- ❌Passwords or PINs
- ❌Bank account numbers or card details
- ❌Your National Insurance number
- ❌NHS number or detailed medical records
- ❌Your full address or date of birth together
- ❌Photos of identity documents (passport, driving licence)
The rule of thumb: If you would not read it out loud on a bus, do not type it into AI.
🔍 How to Check What AI Tells You
AI can be wrong. It sometimes makes up facts, names, dates, and even entire books that do not exist. This is called "hallucination." The problem is that AI sounds confident even when it is wrong.
How to verify important information:
- ✅Search for it. Copy a key fact and search for it on Google. If you cannot find it from a trusted source, be suspicious.
- ✅Check official sources. For health, check the NHS website. For money, check GOV.UK or your bank. For law, check Citizens Advice.
- ✅Ask a real person. If AI gives you medical, legal, or financial advice, always run it past a professional.
- ✅Ask AI to show its sources. You can type "Where did you get that information?" It may not always have a source, and that is a warning sign.
- ✅Be extra careful with numbers. AI often gets statistics, dates, and phone numbers wrong.
🎣 AI-Powered Scams to Watch For
Scammers are using AI to make their tricks more convincing. Here are the main ones to watch for:
📧 Better fake emails
Scam emails used to be easy to spot because of poor grammar. Now scammers use AI to write perfect English. Look for other warning signs: unexpected requests for money, urgency ("act now!"), or emails from addresses that do not match the company they claim to be from.
🗣️ Fake voice calls
AI can now clone someone's voice from just a short recording. Scammers have used this to call people pretending to be a grandchild or family member in trouble, asking for money. If you get a distressing call asking for money, hang up and call the person back on their real number.
🖼️ Fake images and videos
AI can create realistic photos of people who do not exist or put real people into fake situations. Be wary of surprising images or videos on social media, especially if they are trying to make you feel angry or scared.
🌐 Fake AI websites
Some websites pretend to be AI tools but are really designed to steal your information. Stick to the well-known ones: ChatGPT (chat.openai.com), Claude (claude.ai), and Google Gemini (gemini.google.com). Be careful with AI tools you find through ads or pop-ups.
💰 "AI investment" schemes
Be very wary of anyone promising guaranteed returns from "AI trading" or "AI investments." These are almost always scams. No legitimate investment is risk-free.
🎭 How to Spot a Deepfake
Deepfakes are getting more realistic, but there are still signs to look for:
- 👀Watch the eyes. In many deepfakes, the blinking looks unnatural or the eyes do not quite focus properly.
- 👄Check the mouth. The lip movements may not perfectly match the words being spoken.
- 💡Look at lighting. The lighting on the face may not match the lighting in the rest of the scene.
- 🖼️Check the edges. Hair and the edges of the face sometimes look blurry or strange.
- 🤔Ask yourself: is this surprising? If a video shows a politician or celebrity saying something shocking, check if trusted news outlets are reporting it too. If only social media has it, be very sceptical.
⚙️ Privacy Settings Worth Checking
Most AI services let you control how your data is used. Here are some settings worth checking:
- 🔧ChatGPT: Go to Settings, then Data Controls. You can turn off "Chat history & training" so your conversations are not used to improve the AI.
- 🔧Claude: By default, Claude does not use your conversations for training on the free tier in most regions. Check their privacy page if you want to be sure.
- 🔧Google Gemini: Go to your Google Activity settings. You can turn off Gemini Apps Activity to stop Google saving your conversations.
📋 Quick Safety Checklist
- ☑️I will not share passwords or bank details with AI
- ☑️I will double-check important facts from AI with trusted sources
- ☑️I will not click links in unexpected emails, even if they look professional
- ☑️If a family member calls asking for money urgently, I will hang up and call them back on their real number
- ☑️I will stick to well-known AI services and avoid unknown ones
- ☑️I will be sceptical of shocking videos or images on social media
- ☑️I will check my privacy settings on any AI service I use
Stay Smart, Stay Safe
The same common sense that keeps you safe in everyday life works with AI too. Be curious, be cautious, and do not be afraid to ask for help. If you are ready to get started, visit our Try It Yourself page.
🧩 Quick Quiz: Staying Safe
No pressure. Just a fun way to check what you have picked up.
Should you share your password with an AI chatbot?