
The face of modern China
Last week a Dutch security researcher discovered that SenseNets, a facial recognition company that worked for the Chinese police, has been monitoring the GPS coordinates of almost 2.6 million people in Xinjiang, the northwestern area of China home to 12 million Uighurs and other indigenous Muslims.
Xinjiang has become a major testing ground for integrating facial recognition technology into the state security apparatus. In the region’s capital of Urumqi, checkpoints and facial scanners are commonplace at transit hubs, hotels and banks.
In shops, kitchen knives on sale are embedded with QR codes that are linked back to the buyers’ identity cards. The government says the measures are needed after a series of terrorist attacks carried out in recent years by Uighur extremists.
The GPS data, collected by the Shenzhen-based firm SenseNets, was publicly accessible for months, according to the researcher, Victor Gevers. Names, ID numbers, addresses, photos, employer details and other personal data was attached to logs of location informations.
(Since being notified, SenseNets has switched off access to the data although it hasn’t commented on the oversights that made it publicly available in the firstplace.)
“The government’s using this technology [facial recognition] to catch people who are considered threats to social stability,” a human rights activist told America’s NPR. “Are they using it to catch thieves? Yes. But it’s mostly used to maintain stability.”
SenseNets (not to be confused with facial recognition giant SenseTime) also provides surveillance support for police in other provinces.
Revenues for its parent company NetPosa, which is listed on the Shenzhen stock exchange, have grown rapidly and its latest filing shows that it is forecasting net profit of up to Rmb480 million for 2018 ($71.4 million).
In an effort to enhance domestic security, government agencies are now the biggest consumers of facial recognition technology in China. In Shenzhen the technology is used to identify jaywalkers. In Beijing, it has even been deployed to catch thieves of toilet paper from public restrooms.
The massive demand for facial recognition means that the artificial intelligence-powered providers have grown into some of the most active suppliers in the world.
At the forefront of the sector are two major startups, SenseTime and Megvii. Beijing-based Megvii was valued at $3.5 billion during its latest funding round last December, with investors that include Alibaba, Ant Financial, Legend Group, Foxconn and state-owned behemoth China Construction Bank.
Founded in 2011, Megvii operates one of the largest open platforms providing facial recognition, Face++, and the face-based identity verification system, FaceID. It claims that its FaceID platform has authenticated the identities of 400 million people worldwide and its technology also powers the “Smile to Pay” function on Alipay, the electronic payment system that underpins Alibaba’s e-commerce platforms, allowing users to authorise payments with facial scanning.
The government is also a valuable client. Megvii’s technology has been deployed in more than 260 cities in China and has helped police to arrest over 10,000 people. In fact, surveillance like this is its biggest revenue generator, with the Chinese government becoming its largest client, says JRJ.com, a finance portal.
Megvii also counts smartphone firms Oppo, Vivo and Xiaomi on its list of clients. All three also source facial recognition technology from SenseTime for some of their handsets (perhaps China’s smartphone firms will want to avoid either of Megvii or SenseTime becoming too dominant).
“The near-insane quality control and technical requirements have put us under enormous pressure,” Megvii Vice President Wu Wenzhou told Deep Web, a Chinese site on AI, about a recent collaboration with Vivo for a flagship handset targeting customers in the India market. “But Vivo insisted that we lower the time it takes to identify a user from four milliseconds to just one millisecond. Vivo says, why do you want to set a ceiling for yourself?”
Meanwhile, Megvii is rumoured to be mulling an initial public offering in Hong Kong. The company, which claims that it has been profitable since 2017, is said to want to raise at least $1 billion. Archrival SenseTime is also said to be looking into an IPO with the goal of raising $2 billion.
Alibaba has been spreading its bets across the sector: the e-commerce giant is an investor in three out of the four facial recognition unicorns – namely, Megvii, SenseTime, CloudWalk and Yitu.
Caijing has reported that the e-commerce giant poured Rmb1.5 billion into the most recent fundraising round for SenseTime last April (see WiC405), while the Jack Ma-backed private equity fund Yunfeng is a backer of Yitu.
“The reason for the overlapping investments is to strengthen the technological gaps. Compared with other tech giants, Alibaba’s strength in facial recognition doesn’t appear to be very outstanding. What it is hoping to do is to recruit the top engineers in the industry to enhance its competitiveness,” ventures The Next Unicorn, a tech portal. “More importantly, it is a defensive move against Tencent. When it comes to artificial intelligence, Tencent’s Youtu and AI Lab [two of its AI subsidiaries] are already working closely with businesses in sectors like e-commerce. So, of course, Alibaba can’t afford to lose out on that front.”
© ChinTell Ltd. All rights reserved.
Sponsored by HSBC.
The Week in China website and the weekly magazine publications are owned
and maintained by ChinTell Limited, Hong Kong. Neither HSBC nor any member of the HSBC group of companies ("HSBC") endorses the contents and/or is
involved in selecting, creating or editing the contents of the Week in China website or the Week in China magazine. The views expressed in these
publications are solely the views of ChinTell Limited and do not necessarily reflect the views or investment ideas of HSBC. No responsibility will
therefore be assumed by HSBC for the contents of these publications or for the errors or omissions therein.