“Responding to media inquiries, Megvii has asserted that it has no agreement for the use of its technology in the Chinese police app, and indeed that it has no knowledge of how its technology was integrated into the app.”
The surveillance state that Chinese authorities have built in the country’s Xinjiang region is more intensive than previously thought, with technology from a major Chinese biometrics specialist playing a key role, according to a new report from Human Rights Watch.
China’s harsh surveillance of the large population of Uyghur muslims living in Xinjiang was already well-known. But Human Rights Watch’s new analysis of a mobile app used by Chinese police in the region, which was reverse-engineered by the Berlin-based cybersecurity firm Cure53, reveals that police are logging extensive personal data about residents including things like blood type, political affiliation, and ‘suspicious activities’ such as collecting money from a mosque or using a door other than the front door to enter and exit one’s home.
The analysis also found that the police app is designed to let authorities collect face biometrics data using technology from Face++, a company owned by Megvii. An expert in AI-driven facial recognition, Megvii has reportedly raised hundreds of millions of dollars from prominent investors like Alibaba, and the company announced last September that it had teamed up with Himax and MediaTek to develop 3D face modelling technology for Android mobile devices.
Responding to media inquiries, Megvii has asserted that it has no agreement for the use of its technology in the Chinese police app, and indeed that it has no knowledge of how its technology was integrated into the app.
Human Rights Watch’s report comes soon after a security researcher claimed in March that SenseNets, another provider of facial recognition technology for China’s state surveillance apparatus, was using technology provided by Microsoft. Like Megvii, Microsoft denied knowledge of how its technology had wound up in the repressive surveillance system.
(Originally posted on FindBiometrics)