A phone call in your mother's voice. A video of a famous CEO recommending an app. A “police officer” on a video call. Thanks to AI, all of these can now be faked cheaply and convincingly, and AI-powered fraud is in the news almost every week. The good news: the defence is simple, and it works even when the fake is perfect.
What is a deepfake?
A deepfake is media created or altered by AI to show a real person saying or doing something they never did. It uses the same kind of neural networks and image generation you met earlier in this course, trained on a person's photos, videos or voice.
- Face swaps: one person's face placed onto another's body in a video.
- Lip-sync fakes: a real video with the mouth changed to say new words.
- Voice clones: a copy of someone's voice from just a few seconds of audio.
- Fake images: realistic photos of events that never happened.
Not every deepfake is harmful: films, dubbing and accessibility tools use the same technology. The problem is when it's used to deceive.
How AI scams work
1. Collect
A few seconds of your voice from Reels, YouTube or a short phone call.
→2. Clone
Free or cheap AI tools copy the voice, or swap a face in a video.
→3. Pressure
A call or video with an emergency: accident, arrest, a “once-only” investment.
→4. Cash out
Money is demanded fast, by UPI, gift cards or crypto, which is hard to reverse.
The most common AI-powered scams right now:
- Family emergency calls: a cloned voice of a child, parent or sibling who needs money urgently.
- Digital arrest: fake police or CBI on a video call demanding money to “clear your name”.
- Celebrity investment ads: deepfake videos of famous people promoting trading or crypto apps.
- Boss or CEO fraud: a fake video call from a senior manager asking staff to transfer company money.
- Romance and sextortion: fake profiles, or fake explicit images used for blackmail.
- AI-written phishing: perfectly written messages and look-alike websites with no spelling mistakes to give them away.
Test yourself: scam or genuine?
🛡️ Scam or genuine?
6 messages📞 Phone call, unknown number
Your younger brother's voice, crying: “I've had an accident, the police want ₹50,000 or they'll file a case. Send it to this UPI ID now. Please don't tell Mum and Dad.”
🎥 Video call on WhatsApp
A man in uniform says he's a CBI officer. A parcel in your name had drugs in it. You're under “digital arrest”, must stay on camera and transfer your savings for “verification”.
📺 Video ad on social media
A famous business leader looks straight at the camera: “I've invested in this new AI trading app. Put in ₹20,000 today and it's guaranteed to become ₹2 lakh in a month.”
💬 WhatsApp message
“Part-time job! Like YouTube videos from home and earn ₹3,000–5,000 a day. Pay ₹999 to unlock premium tasks.”
🏦 Notification inside your bank's official app
“Scheduled maintenance: UPI services may be slow on Sunday from 1 am to 3 am.” No link, no request for details.
📧 Email from your college's usual address
“Fest volunteer sign-ups close Friday. Register on the student portal.” It points you to the portal you normally log in to.
How to spot a deepfake
Visual and audio clues still catch many fakes. Use this checklist on anything suspicious:
✅ Red-flag checklist
Got something suspicious? Tick everything you notice.
Nothing ticked yet. Remember: a clean-looking video or voice can still be fake. The message matters more than the pixels.
The STOP rule
Slow down
Scams rely on panic. Any real emergency can survive a two-minute pause.
Think: who is really asking?
Is this actually your relative, your bank, the police? Or just a voice and a face?
Other channel
Hang up and call back on a number you already have, or message them on a different app.
Pay? Never under pressure
No real person or agency needs an OTP, a gift card or an instant UPI transfer to “fix” anything.
Protect your family in 5 minutes
- Set a family safe word. A word only your family knows. If an “emergency” caller can't say it, hang up.
- Talk to parents and grandparents. They are targeted most. Show them this page and the STOP rule.
- Lock down your voice and face. Make social accounts private where you can, and think before posting long videos of yourself talking.
- Never share OTPs, PINs or passwords, whoever asks. Banks and police never need them.
- Turn on two-factor authentication on email, WhatsApp and banking apps.
What to do if you're targeted
- Stop contact immediately and don't send more money, even if they threaten you.
- Call your bank right away if you paid: quick action can sometimes freeze the transfer.
- Report it. In India: call 1930 or file at cybercrime.gov.in. In the US: reportfraud.ftc.gov.
- Keep evidence: screenshots, phone numbers, UPI IDs and transaction references.
- Warn others in your family and friend groups: scammers often try the same trick again.
Where to go next
You know how AI can be misused. Now build something useful with it: the final part of the course is build your first AI chatbot.
Key takeaways
- ✓Deepfakes are AI-made videos, images or voices of real people.
- ✓AI scams use fake emergencies, fake authority and fake celebrity tips to rush you into paying.
- ✓Visual glitches help, but the best fakes have none: judge the request, not the pixels.
- ✓Use STOP: Slow down, Think who, verify on an Other channel, never Pay under pressure.
- ✓Set a family safe word, never share OTPs, and report fraud fast (1930 / cybercrime.gov.in in India).
Frequently asked questions
What is a deepfake?
A deepfake is a video, image or audio clip made or edited with AI so that it shows a real person saying or doing something they never did. Modern tools can copy a face from photos and a voice from a few seconds of audio, which makes fakes increasingly hard to spot by eye or ear alone.
How do AI voice cloning scams work?
Scammers collect a short clip of someone's voice, often from social media videos, and use an AI tool to clone it. They then call a family member with a fake emergency, such as an accident or arrest, and demand money quickly through UPI, gift cards or crypto, often asking the victim to keep it secret. The fix is to hang up and call the person back on a number you already know.
How can I tell if a video is a deepfake?
Look for lips out of sync, unnatural blinking, blurry edges around hair or glasses, lighting that doesn't match, and glitches when the person turns or covers their face. But these signs are getting rarer as tools improve, so the most reliable checks are about context: who is asking, what they want, and whether you can confirm it through another trusted channel.
What is a digital arrest scam?
A digital arrest scam is a fraud, common in India, where criminals pose as police, CBI, customs or other officials on a video call. They claim you are linked to a crime, tell you that you are under “digital arrest” and must stay on camera, and pressure you to transfer money to clear your name. Real agencies never arrest people over video calls or ask for money this way.
What should I do if I've been targeted by an AI scam?
Stop all contact and don't send more money. If you already paid, call your bank immediately to try to block the transaction. In India, call the national cybercrime helpline 1930 or report at cybercrime.gov.in; in the US, report at reportfraud.ftc.gov. Save screenshots, numbers and transaction IDs, and warn family members who might be targeted next.
🧠 Quick check: deepfakes and AI scams
1. You get a call in your friend's voice asking for urgent money by UPI. What's the best first step?
2. Which is the strongest sign of a scam?
3. Why can't you rely only on spotting visual glitches in deepfakes?
4. What does the “O” in the STOP rule stand for?
Checking your access…