Liveness Detection: Telling Humans Apart from Deepfakes

With just a free app, anyone can now create AI-generated content that looks incredibly real. You can edit your photo to make it look like you’re posing with a celebrity, or blend your childhood picture with your current face. At a glance, these AI-generated images (often called deepfakes) seem fun, right?

But what if someone uses AI to fake your ID photo and opens an online loan account in your name without you ever knowing?

That’s the problem. Deepfakes are no longer just for entertainment. They’re now being used as tools for fraud.

In Indonesia, cybercrime has surged dramatically. According to a 2024 report by BSSN, there were more than 403 million anomalous cyberattack traffic events in a year—many of them exploiting weak verification systems that can’t distinguish between real humans and deepfakes.

There is a solution: using liveness detection in identity verification systems.

If you’ve ever registered for a bank app, online loan, or paylater service and were asked to take a selfie while turning your head or blinking, that’s liveness detection in action. So how does this technology protect your face from deepfake fraud? Let’s dive in.

What Is Liveness Detection?

Liveness detection is a technology used to ensure that the person undergoing identity verification is a real, live human—not a photo, a video, a mask, or a deepfake.

What’s the difference compared to regular face verification?

• Standard face verification only checks if your face matches the photo on your ID.

• Liveness detection goes further. It checks whether the face is alive, present in real time, and not something generated or replayed by AI or software.

Liveness detection is crucial in identity verification processes, especially to prevent fraud. Cases of fake faces and synthetic identities are already emerging in Indonesia, especially in online lending and digital financial services.

Examples include:

• Fake faces used to open online loan accounts and withdraw funds.

• Deepfake videos used to impersonate customer service agents or executives.

• Stolen identities used for money laundering or credit manipulation.

VIDA’s survey even shows that 90% of business professionals in Indonesia don’t know how to protect their companies against deepfake threats.

How Liveness Detection Works

How does liveness detection know whether it’s looking at a real human?

The technology analyzes biological responses and natural human patterns, such as:

• eye movements

• facial depth

• micro-expressions

• natural skin texture and light reflection

These are characteristics of real human faces that deepfakes, screens, printed photos, or masks can’t perfectly replicate.

When liveness detection detects a fake face, the system automatically rejects the verification attempt. Some of the red flags include:

• a face shown on another screen (replay attack)

• flat, 2D images with no depth

• unnatural or static lighting

• unrealistic skin texture (deepfake artifacts or silicone masks)

Where Is Liveness Detection Used?

Chances are, you’ve already used liveness detection without realizing it. Let’s look at some everyday scenarios:

1. Fintech & Online Lending

Liveness detection is critical for online loan applications. You’ve probably heard of illegal loan apps using other people’s ID cards to apply for loans. With liveness detection, the system doesn’t just match your ID photo and selfie—it confirms that the person in front of the camera is alive and present.

So even if someone steals your ID, they still can’t pass onboarding without your real, live face.

2. Digital Banking

Many banks and digital banks now use liveness detection during onboarding—and even for login. In some countries like Singapore, India, and Malaysia, banks are moving away from passwords and OTPs. Customers simply log in using facial recognition.

But if it’s just face matching, fraudsters could use a printed photo or video. That’s why liveness detection is layered on top—to ensure only a real human face can access the account.

3. Investment & Digital Insurance

When you register for mutual funds, stock trading, or digital insurance, you’re usually asked to submit an ID photo and a selfie.

During that selfie step, liveness detection is often running in the background—to ensure the person creating or accessing the account is truly you, not someone else using your stolen data. Without liveness detection, someone could use your identity to open accounts or claim funds in your name.

4. E-commerce & Logistics

For high-value deliveries like electronics, jewelry, or legal documents, some e-commerce platforms and logistics providers now use liveness detection to verify couriers and recipients.

This ensures that the person delivering or receiving the package is truly authorized—not just someone pretending to be.

What Happens If Identity Isn’t Protected by Liveness Detection?

Without liveness detection, digital identity systems are extremely vulnerable. Many platforms still rely on photos, passwords, or OTPs—everything that can be copied, stolen, or hacked.

Here are some real risks:

1. Deepfakes Can Pass Verification

If a system can’t distinguish between real and AI-manipulated faces, a deepfake can be treated as a legitimate user.

2. Identity Theft via Fake IDs

Imagine someone creates an AI-generated “ID card” and uploads it to a financial platform. Without liveness detection, the system might accept it, even if the face is synthetic.

3. Account Takeovers

OTP codes can be stolen through SIM swap, phishing, or malware. Without liveness as an extra layer, hackers can take over accounts using stolen OTPs.

4. Brand and Business Reputation Damage

If a platform allows fraudsters to slip through weak verification, user trust will plummet—and the cost of remediation can be huge.

VIDA’s Liveness Detection

VIDA provides liveness detection technology that is iBeta Level 2 PAD certified—one of the world’s highest anti-spoofing standards.

Key Features:

• ColorFlash: Dynamic lighting technology that reads how real light behaves on human skin, helping distinguish live faces from manipulated media.

• Real-time deepfake detection: Detects manipulated faces within seconds.

• Full integration with KYC and FaceToken: Ensures biometric-based identity verification from onboarding to authentication.

• Optimized user experience: Just one selfie—no complicated steps—while the system does the checks in the background.

• ConversionOps: Human-in-the-loop system to reduce false rejects and save genuine user onboarding.

• AI Shield & Fraud Detection: Identifies fraud patterns before losses occur using machine learning and behavioral analytics.

Facing the growing threat of deepfake-enabled fraud, VIDA’s liveness detection is a practical, proven solution to detect and stop attacks before they cause damage.