Blog

What the attack surface actually looks like

Detection benchmarks, what synthetic identities reveal about how vision models fail, and how the data behind them gets built. Sourced and dated — every external claim links to where it came from.

4 August 2026· 3 min read

What Is Synthetic Biometric Data?

Synthetic biometric data is AI-generated face, voice, or fingerprint data built to train and test identity systems without using a real person's identity. Definition, how it's made, and why it exists.

4 August 2026· 3 min read

Why Detectors Need to Train Against Their Own Weapons

Group-IB tracked 8,065 deepfake biometric injection attacks against loan-application liveness checks. How deepfakes bypass liveness detection, and why detection models need adversarial synthetic data to hold up.

4 August 2026· 3 min read

Synthetic Data Is a Small Market. The AI-Compute Cycle It Feeds Isn't.

Synthetic data market size forecasts for 2030 range from $1.8 billion to $9.7 billion depending on scope. Why the AI-compute cycle it drives, projected at up to $1.7 trillion, matters more to infrastructure owners.

4 August 2026· 3 min read

The Synthetic Biometric Data Landscape: Who's Building What

A guide to the synthetic biometric data category: Precise Biometrics, Datagen, Synthesis AI, Mostly AI, and TessLabs, and where each one actually sits in the identity-verification stack.

4 August 2026· 2 min read

The Compliance Gate AI Regulation Just Opened

A fraudster opened 46 bank accounts with stolen IDs and deepfake video. Why eIDAS 2.0 and sovereign-AI procurement now require biometric assurance platforms don't yet have.

4 August 2026· 4 min read

The $25 Million Zoom Call

A finance employee wired $25 million after a video call where every participant, including the CFO, was a deepfake. Updated with 2026 cases. Why payment authentication needs detection trained on realistic synthetic identities.

4 August 2026· 2 min read

46 Accounts, One Face: The Digital Identity Problem

One fraudster opened 46 bank accounts with stolen IDs and deepfake video. Why national digital identity programs need synthetic testing data, not more real citizen biometrics.

4 August 2026· 3 min read

The Interview That Wasn't: Deepfakes and the New Hiring Attack Surface

North Korean operatives are passing live video interviews with real-time deepfake filters. Why frontier-model identity checks need synthetic data to catch what they miss.

4 August 2026· 4 min read

The Frontier-Model Onboarding Gate Nobody's Guarding

AI platform KYC verification is now mandatory at Anthropic and OpenAI. Why frontier-model identity checks face the same deepfake exposure that's already defeating bank KYC, and what closes the gap.

4 August 2026· 3 min read

Can Your Robot Tell a Human From a Projection?

Security researchers scored commercial robots as low as 39.9% on face-recognition and liveness maturity. Why humanoid systems need anti-spoofing built in, not bolted on.

4 August 2026· 3 min read

eIDAS 2.0 and the Identity-Assurance Gate Most Platforms Aren't Ready For

eIDAS 2.0 requirements and compliance checklist: LoA High verification, relying-party registration, the January 2027 deadline, and why biometric assurance is the gate most platforms haven't tested.

4 August 2026· 3 min read

Deepfaked Command: The Battlefield Nobody's Guarding

A fabricated Pentagon explosion image became a DHS case study on adversarial AI. Why defense and intelligence need generate-and-detect capability, not detection alone. Updated with the DoD's Hive AI and Jericho Security contracts.

4 August 2026· 2 min read

When the Prime Minister's Voice Isn't the Prime Minister's

82 deepfakes targeted public figures across 38 countries in a year. Why platform integrity now depends on detection that works regardless of how the fake was made.

4 August 2026· 3 min read

How a $20 Deepfake Toolkit Beats Bank-Grade Liveness Checks

How deepfakes bypass KYC verification and liveness checks: virtual camera injection, off-the-shelf toolkits under $20, and why deepfakes now account for 11% of global fraud.

4 August 2026· 3 min read

Deepfake Detection Vendors in 2026: What Each One Actually Catches

A comparison of deepfake detection vendors by modality and use case: iProov, Sumsub, Onfido, Incode, Pindrop, Sensity, Hive, and where synthetic training data fits underneath them.

4 August 2026· 3 min read

The Deepfake Detection Market Is Growing 43-48% a Year. Here's Why Watermarking Won't Save It.

Deepfake detection market size estimates for 2030 range from $1.7 billion to $8 billion depending on scope, growing at 42-48% CAGR. Why watermarking doesn't solve the underlying problem.

From reading to measuring

Name the model that needs fixing

The first step is a sample you score on your own detectors — no cost, no commitment.