Chinese government facial recognition system hacked by tax fraudsters

Occurred: 2018-2021

Two men in China tricked the country's State Taxation Administration's facial recognition identity verification system into creating fake invoices, raising fears about the stength of its security. 

Two fraudsters, identified only by their surnames Wu and Zhou, manipulated high-definition photographs with an app available on the black market that turns photos into videos to create deepfake videos that simulated facial movements like nodding, blinking and mouth opening.

They then used a smartphone to bypass the government system camera during facial authentication, instead feeding their doctored videos into and successfully fooling the identity verification process.

The two then created fake tax invoices, defrauding the taxation department by 500 million yuan (approximately USD 76.2 million). 

The fraudsters started their operation in 2018 and have since been captured and prosecuted in Shanghai.  

System 🌐

Operator: State Taxation Administration
Developer: State Taxation Administration
Country: China
Sector: Govt - finance
Purpose: Verify identity
Technology: Facial recognition; Deepfake - image; Deepfake - video; Generative adversarial network (GAN); Neural network; Deep learning; Machine learning
Issue: Security

Page info
Type: Incident
Published: March 2021
Last updated: February 2024