AI Explorers · lesson 20 of 24 · 25 min
A video goes around your group chat of a famous rapper saying something wild. It looks real. It sounds real. It isn’t. Welcome to the age of deepfakes.
Your data feeds AI
AI systems need data, and a lot of that data comes from people. Every search, like, location check-in, voice command and photo can become data. Privacy is about having control over your own personal information: who collects it, how it is used, and who it is shared with.
Some questions to ask about any app:
- What data does it collect? Location, contacts, photos, voice?
- Does it need that data to work, or is it extra?
- Can I turn some of it off in settings?
- Who else gets the data?
In the US, a law called COPPA requires apps and websites aimed at kids under 13 to get a parent’s permission before collecting their personal information. That is one reason many apps have age limits.
Deepfakes
A deepfake is fake video, audio or image content made with AI to look or sound like a real person. Generative AI can now copy someone’s face or voice from samples of them.
Deepfakes can be used for fun, like movie effects or a clearly labeled joke. They can also cause real harm:
- Spreading lies about a public figure.
- Scams, like a fake voice call that sounds like a family member asking for money.
- Bullying, by putting someone’s face into embarrassing content.
Making a deepfake of a real person without their permission can hurt them badly, and in some places it is against the law. Never create or share fake content of classmates or anyone else.
How to spot fakes
No single trick catches every fake, and fakes keep getting better. Use your judgment and these habits:
- Check the source: who posted it first? Is it a trusted news outlet or an account you’ve never heard of?
- Look for other reporting: if something shocking is real, several reliable sources will usually report it.
- Notice details: odd blinking, blurry edges around the face, mouth not matching words, strange hands, or a voice that sounds flat.
- Pause before sharing: content designed to make you angry or shocked spreads fastest. Slow down.
- Make a family code word: agree on a secret word your family can use to confirm a real emergency call.
Privacy also includes other people’s information. Posting a photo of a friend, sharing a screenshot of a group chat, or recording someone without asking can put their information out there without their permission. A good rule: ask before you post or share anything about someone else, and respect their answer.
Try it
- Choose one app you or your family use often.
- With a parent, open its privacy or permissions settings and list what it can access (location, camera, microphone, contacts).
- For each permission, decide: needed, nice to have, or not needed. Turn off anything not needed, if your parent agrees.
- Then write a 5-step “Before I Share” checklist for your family group chat, using the spot-a-fake habits above.
- Share your checklist with your family and agree on a family code word.
Check yourself
What is a deepfake?
Fake video, audio or images made with AI to look or sound like a real person.
Name one good habit for spotting fakes.
Check the original source, look for reliable reporting, notice odd details, or pause before sharing.
What does COPPA require?
In the US, apps and sites aimed at kids under 13 need a parent’s permission before collecting their personal information.
Key words
- Privacy: control over your personal information.
- Personal information: details that identify you, like your name, address or face.
- Deepfake: AI-made fake media of a real person.
- Permission: access an app is allowed to have on your device.
Previous lesson: Bias and fairness in AI
Back to AI Explorers
Next lesson: AI, jobs and who gets to build it