Deepfake Scams in Singapore: How AI Voice and Video Fraud Works — and How to Fight Back

Deepfake Scams Are Here — And Singapore Is a Prime Target

Deepfake technology has evolved from a novelty to a serious financial threat. In one high-profile case, scammers used AI to recreate a CFO's face and voice during a Zoom call, convincing an employee to transfer US$499,000 (over S$670,000) to fraudulent accounts. The entire video conference was fake — every participant was a deepfake.

In March 2025, the Singapore Police Force (SPF), Monetary Authority of Singapore (MAS), and Cyber Security Agency (CSA) issued a joint advisory warning about the rise of deepfake-enabled scams. According to recent surveys, 56% of Singapore businesses have already reported audio deepfake incidents.

---

Types of Deepfake Scams in Singapore

Type	How It Works
Deepfake video calls	Scammers recreate your boss's or colleague's face and voice on a video call to authorize urgent wire transfers
AI voice cloning	Clone a family member's voice using short audio samples, then make fake emergency calls demanding money
Investment scam videos	Create fake endorsement videos of celebrities or financial figures promoting fraudulent investment schemes
Romance scams	Use deepfake video to build trust in online relationships before requesting money

---

How to Detect Deepfakes

Deepfakes are improving rapidly, but current technology still has telltale weaknesses:

Visual Cues

• Ask the person to turn sideways — deepfakes struggle with profile views and head rotation
• Watch for unnatural blinking — AI often gets eye movement wrong
• Check lip sync — audio and lip movements may be slightly out of sync
• Look for background inconsistencies — lighting, shadows, or background elements may flicker

Behavioural Cues

• Ask unexpected questions — questions the real person would know but a scammer wouldn't
• Request a specific gesture — "Can you wave with your left hand?" or "Touch your nose"
• Listen for audio artefacts — robotic undertones, strange pauses, or background noise changes

---

Verification Strategies for Financial Transactions

Strategy	Implementation
Dual approval	Require two people to authorize any financial transfer above a threshold
Callback verification	After receiving a request, hang up and call back on a known, verified number
Secret code words	Establish pre-agreed verification phrases with family and key colleagues
Multi-channel confirmation	Verify requests through a different communication channel than the one used for the request

---

Share Verification Passphrases Securely

One of the most effective defences against deepfake scams is establishing secret verification phrases with your family and close colleagues. Before any video call requesting money or sensitive action, verify identity using the shared secret phrase.

But where do you store this passphrase securely? Sending it via iMessage or Messenger defeats the purpose — if someone has compromised your communications, they could intercept it.

Use LOCK.PUB to create a password-protected memo containing your family's verification passphrase. Share the LOCK.PUB link once, in person, with each family member. If someone ever calls claiming to be a family member in an emergency, you can verify them by asking for the passphrase — something a deepfake scammer would never know.

---

What Businesses Should Do

1. Implement verification protocols for all financial transactions over a threshold
2. Train employees to recognize deepfake red flags
3. Establish callback procedures — never act on video call instructions alone for wire transfers
4. Use multi-factor authentication for authorizing payments
5. Create an internal reporting channel for suspected deepfake encounters

---

What Individuals Should Do

1. Establish family code words for emergency verification
2. Be sceptical of urgent video calls requesting money or sensitive actions
3. Verify through a second channel — if your "boss" calls on Zoom, call them back on their personal phone
4. Limit voice/video content online — the less audio/video available, the harder it is to create convincing deepfakes
5. Stay informed — follow SPF, MAS, and CSA advisories

---

The Future of Deepfake Defence

As AI improves, deepfakes will become harder to detect visually. The best defence will shift from "spotting the fake" to "verifying the real" — using pre-established verification systems that no amount of AI can replicate.

Start building your verification system today. Create a secret passphrase with your family on LOCK.PUB and agree on a verification protocol before you need it.

Protect yourself at LOCK.PUB — create a secure verification memo today.