Everyday AI · Safety & Security
Spot AI scams, deepfakes and cloned voices
Did you receive a suspicious video, voice note, or emergency money request that seems real?
Quick answer
Treat any urgent money request by voice note, video or chat as unverified, even if it sounds exactly like family. Do not send money or share codes. End the call, phone the person back on the number you already saved, and ask a private question only they can answer. Keep screenshots if you plan to report it.
Listen to this guide
AI narration · English · Prompt text stays on the page
Try ElevenLabs ↗ · Affiliate link
What to do
- Stop before sending money: AI voice cloning needs only a 3-second audio clip from social media to imitate a family member.
- Check the lips and teeth in videos: Deepfakes often have blurred teeth, unnatural eye blinking, or audio that lags slightly behind lip movement.
- Call the person back on their real, known mobile number—not the caller who just contacted you or the link in the message.
- Ask a private family offline question that only the real person knows (e.g., "Anong pangalan ng una nating aso?").
Try this prompt
Paste the suspicious text or description of the video/audio to check for common scam patterns.
Help me assess a suspicious voice/video/message without claiming AI can prove whether it is real or fake from appearance alone.
WHAT I HAVE
Type: [VOICE CALL / AUDIO / VIDEO / VIDEO CALL / MESSAGE]
What the person/account claims: [FACTS]
What they ask me to do: [PAY / SEND CODE / MOVE MONEY / URGENT ACTION / OTHER]
Observable clues: [ONLY WHAT I PERSONALLY NOTICED]
Relationship/person being impersonated: [GENERAL RELATIONSHIP / PUBLIC FIGURE / UNKNOWN]
RULES
- Do not declare media authentic or AI-generated based only on visual/audio artifacts.
- Treat urgency, secrecy, payment requests, credential/code requests and unusual contact methods as behavioral evidence separate from “deepfake detection.”
- Do not reproduce or request passwords, OTPs, account numbers or other secrets.
- Do not recommend calling a number or using a link supplied by the suspicious contact.
- Do not encourage confronting the suspected impersonator.
OUTPUT
1. Claims/requests made.
2. Observable risk signals, with evidence.
3. What cannot be determined from the media alone.
4. Independent identity-verification plan: use a known number/channel, shared family verification phrase if one already exists, or another trusted contact.
5. Immediate do-not-do list.
6. If money/credentials were already shared, safe next steps using independently reached official channels.
7. Evidence-preservation checklist that avoids forwarding harmful links/files unnecessarily.
End with: “Verify the person through a separate trusted channel before acting; AI-generated-media detection is not proof.”See a sample result
Illustrative example. Your answer may differ.
Input: A Facebook video showing a famous doctor endorsing an unregistered herbal pill with robotic lip movements. Result: Identifies unverified endorsement red flags, advises checking FDA advisory search, and warns never to buy medicine from unverified social ads.
Before you use it: No bank, e-wallet, or genuine family emergency requires you to send money within 5 minutes without phone verification.
Philippine notes
- Report scams to the PNP Anti-Cybercrime Group (ACG) hotline or through your bank/e-wallet app immediately.
- GCash and banks will never call or message to ask for your MPIN or OTP.
More everyday guides
Did you receive an unsolicited online job offer or work-from-home task and wonder if it is a scam?
Did you see a GCash or Maya transaction you did not authorize and need a clear timeline and support message?
Did you get a courier or SMS message that asks for a fee, a click, or personal details and want a careful red-flag check?
Do you want AI to help with a bill, receipt or money question?