Facial Recognition in China

Written: April  2023

Fortune Magazine published an article about facial recognition in China. The title was “Every Step You Take, They’ll Be Watching You”.  You might think with a population of 1.4 billion, it would be hard to conduct surveillance over such a large numbers of people. It actually is not hard because Chinese tech companies are very focused on facial recognition. Fortune wrote about three leading facial recognition companies. The largest is SenseTime, which started out as a research project at the University of Hong Kong. SenseTime is now the world’s most valuable AI startup, with a valuation exceeding $91 billion.  In second place, at an estimated $4 billion valuation, is Megvii.  Megvii created the world’s largest open-source facial recognition platform called Face++. More than 300,000 developers use Face++ to build their own face detection programs. A third but small company is Intellifusion, which provides facial recognition software enabling police to gain the identity of pedestrians who jaywalk. Collectively, the three companies have the most powerful AI facial recognition systems in the world.

What is unique about the use of facial recognition in China is how they think about it. Fortune’s article summed it up this way,

Founded in 2011 by three graduates of Tsinghua University—China’s MIT equivalent—Megvii claims it wants to “build the eyes and the brain” of Chinese cities and extend police powers to a point “beyond what is humanly possible.”  For China’s government, that means not only being able to identify any of its 1.4 billion citizens within a matter of seconds but also having the ability to record an individual’s behavior to predict who might become a threat.

Chinese facial recognition companies see huge potential in selling their technology to the local and national Chinese governments but take no responsibility for privacy considerations. Megvii has created technology to support a nationwide surveillance program called Skynet. Skynet is a Chinese government surveillance system which has installed more than 20 million cameras in public areas across the country.  The Chinese government says the Skynet AI-enhanced cameras are “the eyes that safeguard China”, and they have enabled the apprehension of approximately 2,000 suspects. 

The potential for major privacy issues and misuse of the facial recognition technology is significant.  One surveillance system can determine if a suspect moves more than 1,000 feet from where they live. If a suspect escapes from view, and he or she later disappears into the crowds, the surveillance system can find their locations within minutes. These approaches could be applied to dissidents or minorities as easily as suspects or criminals. The technology companies say policing government abuse is not their job.

Chinese surveillance giant Hikvision has repeatedly denied reports the company is complicit in human rights abuses targeting Uyghurs, one of China’s 55 officially recognized ethnic minorities, in China’s northwestern region of Xinjiang. But an internal review of the company’s contracts with police agencies in the region reveals the company has known since at least 2020 that some of its Xinjiang contracts were a “problem,” Axios’ Bethany Allen-Ebrahimian and Ina Fried reported. The contracts included language about targeting Uyghurs as a group, according to a recording of a recent private company meeting obtained by technology trade publication IPVM and shared exclusively with Axios.

Axios said the Chinese government is perpetrating an ongoing campaign of genocide and mass detention of Uyghurs and other ethnic minorities in Xinjiang. Procurement documents reportedly show Hikvision cameras have been installed in public spaces across Xinjiang and in mass detention facilities, and Hikvision cameras have captured footage which has led to the detention of Uyghurs. Hikvision has also advertised it offers biometric surveillance technology which can track ethnic minorities, including Uyghurs, though in 2020 the company stated its products no longer offer such capability.

Human rights groups and the U.S. and other governments have accused Hikvision of participating in human rights abuses in Xinjiang — allegations the surveillance giant has rejected. The use of Hikvision’s facial recognition technology is a controversial issue.

There are legitimate concerns about privacy and human rights, but there are also potential benefits to the technology, such as improved public safety and reduced crime. It is important to weigh the pros and cons of the technology before deciding about its use. The need for regulation around the world is very clear. What is not clear is China’s intention.

In the United States, tech companies are beginning to take more responsibility for the privacy risks of facial recognition but are asking for government regulation. The Artificial Intelligence blog at the MIT Technology Review elaborated,

Microsoft President and Chief Legal Officer Brad Smith wrote in a blog post the company is asking Congress to regulate AI-powered face recognition software. “There will always be debates about the details, and the details matter greatly,” Smith wrote. “But a world with vigorous regulation of products that are useful but potentially troubling is better than a world devoid of legal standards.”

Mr. Smith outlined the key questions he recommends policymakers should discuss:

• Should law enforcement’s use of facial recognition be subject to human oversight and controls?
• What types of legal measures can prevent use of facial recognition for racial profiling and other violations of rights, while still permitting beneficial uses of the technology?
• Should the government create processes that afford legal rights to individuals who believe they have been misidentified by a facial recognition system?

At Amazon Web Services, the issue is much closer to the ground than Mr. Smith’s viewpoint. Amazon shareholders have gone on record demanding the company cease selling facial recognition software to government agencies. The letter sent to Amazon was organized by Open Mic, a non-profit organization focused on corporate accountability.  Employees of Amazon have weighed in also. More than 450 employees signed a letter addressed to Jeff Bezos demanding a company named Palantir, which is working on Immigration and Customs Enforcement’s deportation and tracking program, be banned from using Amazon Web Services which offers Rekognition facial recognition technology.  The employees have also asked Amazon to implement employee oversight for ethical decisions related to Amazon technology.

In summary, facial recognition offers benefits but also great threats to privacy. The bottoms up protests add a focus. What is missing is Congress. Like on so many issues, the politicians can not agree on what to do.