Lesson 5 of 5 · 6 min

Using AI Safely and Wisely

Protect your private data, check facts, spot voice-clone and deepfake scams, guide children, and know when to ask a real professional instead.

In this lesson

  • Know which information never to share with AI
  • Check AI answers and spot voice-clone and deepfake scams
  • Know when to ask a professional instead of an AI

What never to share

Treat an AI chat like a conversation in a public place. Many services store conversations, and some use them to improve their systems unless you switch this off in the settings. Even when a company is careful, it is safer to keep sensitive information out completely.

Never type or upload:

  • Passwords, PIN codes or one-time security codes.
  • ID card, passport or national insurance numbers.
  • Bank card numbers, account details or screenshots of your banking app.
  • Other people’s private information, such as their health, address, phone number or photos, without their permission.
  • Confidential work documents, unless your employer has approved the tool.

You can still get help with personal matters. Just remove or replace the details first, so the AI sees the situation but not the person.

Before: “My name is Sara Haddad, I live at 12 Park Road and my account number is 4481… Write a complaint to my bank.”

After: “Write a complaint to my bank about a payment that was taken twice. Write NAME and ACCOUNT NUMBER where my details should go.” You add the real details yourself after copying the text.

Check the facts

AI can sound very sure and still be wrong. Before you rely on an answer, especially one you will share or act on, check it. Ask the assistant for its sources, then open them yourself to see whether they exist and really say what the AI claims. Compare with an official website or a trusted news source. Be extra careful with numbers, dates, names, quotes and anything recent. Some tools, such as Perplexity, show sources next to each answer, which makes checking easier, but you still need to read them.

Voice clones, deepfakes and scams

The same technology that writes friendly emails can also be used by criminals. AI can copy a person’s voice from a short recording and create realistic fake videos, called deepfakes. Scammers use this to pretend to be a relative in trouble, a manager asking for an urgent payment, or a celebrity promoting an “investment”.

Warning signs include:

  • A call or voice message that creates panic and asks for money fast.
  • A request to pay in an unusual way, such as gift cards, crypto or a new bank account.
  • A video where the lips do not match the words, or the face looks too smooth.
  • Pressure to act now and keep it secret.

What to do: hang up and call the person back on the number you already have. Agree on a family “safe word” that only you know. Never send money or codes based on a voice or a video alone.

Children, and when not to rely on AI

Children are curious, and AI tools can help them learn, but they need guidance. Check the minimum age in each tool’s terms, since many are meant for teenagers and adults, sometimes with parental permission. Use the tools together at first, explain that AI can be wrong, and teach children never to share their name, school or photos. Remind them that an AI is not a friend or a person, even when it sounds kind.

Finally, know when to stop. AI can help you prepare questions for a doctor, understand legal words or set up a simple budget. But it should never make important medical, legal or financial decisions for you. It does not know your full situation and it is not responsible if it is wrong. For those decisions, speak to a qualified professional.

Practice

Open any AI assistant you already use and paste the prompt below. Replace [SUSPICIOUS_MESSAGE] with the text of the message, after removing your real name, phone number and any personal details.

Prompt
I received this message and I am not sure if it is a scam: "[SUSPICIOUS_MESSAGE]". Do not ask me for any personal details. List the warning signs you notice, tell me how risky it looks (low, medium or high) and why, and give me 3 safe next steps.
[SUSPICIOUS_MESSAGE]
e.g. a text saying your parcel is on hold and you must pay a small fee through a link
See an example result

Warning signs:
– It creates urgency: "pay within 24 hours or the parcel will be returned".
– It asks for a payment through a link instead of the delivery company's official app or website.
– The sender is an unknown mobile number, not an official name.
– The greeting is general ("Dear customer") and does not mention what you ordered.

Risk level: High. Real delivery companies rarely ask for small fees by text with a link, and this pattern is very common in scams.

Safe next steps:
1. Do not click the link or reply.
2. If you are expecting a parcel, check its status on the official website or app by typing the address yourself.
3. Block the number and report the message to your phone provider or national scam reporting service.

What to notice: The AI explains why each sign matters and gives clear actions without needing your private data. Use it as a second opinion, but if money or accounts are involved, contact the real company directly.

Key takeaways

  • Never share passwords, IDs, bank details or other people's data
  • Check sources; call back on a known number before acting on urgent requests
  • Use AI to prepare, not to decide, on health, legal and money matters