Chinese human rights practices are in the news again. The White House is reportedly weighing sanctions against Chinese officials and companies that are engaged in or facilitating the mass surveillance and detention of Uighurs in the Xinjiang Uighur Autonomous Region (XUAR). Over the past several months, it has become increasingly clear that the Chinese government is conducting widespread efforts to “re-educate” its largely-Muslim Uighur population in XUAR and impose strict controls over that population’s movements and actions, including through the extensive use of facial recognition software (FRS). And now the Trump Administration, which to date has not focused on other states’ human rights practices, seems to have concluded that China’s actions are worthy of condemnation. Though the administration should not be overly sanguine about the effectiveness of making it harder for a few companies to provide FRS to the Chinese government, there is value in putting down a marker that using FRS this way is not acceptable.
Condemning China’s Use of Facial Recognition Software
The raging trade war with China is a big factor in the Trump administration’s appetite for human rights-related sanctions against China, but pressure from Congress may also play a role. Back in May 2018, Sen. Marco Rubio and Rep. Chris Smith sent a letter to Commerce Secretary Wilbur Ross, asking whether U.S. companies were selling surveillance and crime control technology to Chinese security forces for use against the Uighur population. The letter identified the China’s dramatically increased use of surveillance in XUAR and, drawing from human rights and media reports, noted that U.S.-based companies appeared to have supplied DNA sequencers and surveillance software to the Chinese Ministry of Public Security. Rubio and Smith also asked about the Commerce Department’s process of approving exports to the Chinese government under the Export Administration Regulations. (It is not clear whether or how Commerce replied to this letter.) Subsequent reporting identified several additional companies, including Cisco and iFlyTek, that have willingly or inadvertently facilitated Chinese surveillance and policing.
More recently, seventeen members of Congress sent a letter to Secretary of State Mike Pompeo and Treasury Secretary Steven Mnuchin, urging them to apply sanctions under the Global Magnitsky Act, along with other measures, against the Chinese officials overseeing these repressive policies. Their letter also identifies as possible targets of sanctions two Chinese companies that make video surveillance technology reportedly used by the PRC to track the movements of Uighurs.
I have previously written about China’s widespread use of facial recognition technology, and this latest set of developments brings to the fore a question I raised earlier: How will the United States draw the line between permissible and impermissible uses of FRS? In this case, the Chinese are engaged in so many other troubling practices against the Uighurs that the use of FRS to enable those practices is easy to condemn. If the United States proceeds to impose sanctions, there will be at least one data point about when the United States believes that the use of FRS should be off-limits.
Good FRS/Bad FRS
As Shannon Togawa Mercer and I have explained, however, there are a huge range of contexts in which states—including the United States itself—already use FRS or will soon do so. Sometimes those uses will be benign, as where states employ FRS to reunite families separated by war or law enforcement uses FRS to quickly identify a mass shooter. Technologies that states and private actors can use for good or bad ends can be complicated to regulate. This is also in part because the United States wants those technologies to be both available and widely effective when used for good purposes, but worries about high levels of efficacy and access when the tools are being used for nefarious purposes.
To some extent, this looming debate about the appropriate roles for FRS recalls the debate during the past several years over encryption and backdoors. U.S. companies are key (though not exclusive) producers of both FRS and encryption technologies, and the U.S. government has concerns about those companies exporting the technologies to those who would employ them for problematic purposes, such as to conceal crimes or conduct highly intrusive surveillance. Because the U.S. companies are not exclusive producers of encryption or FRS tools, though, there is good reason to think that blocking the sale of those tools will not prohibit foreign actors from finding and employing substitutes. As the Epoch Times notes, “Large state-backed Chinese tech firms are playing a major role in the development of digital police tools.” If China is blocked from purchasing FRS from U.S. companies, it seems likely that Chinese companies soon will be able to take their place. Indeed, in April 2018, a Chinese FRS company called SenseTime became the world’s most valuable artificial intelligence start-up.
Laying Down a Marker
All this is not to say that the United States should not try to make it harder for the Chinese government to gain access to FRS tools, because that government is so likely to use FRS for repressive purposes. Rather, it is to say that the United States should not be overly optimistic that cutting off access to U.S.-developed FRS will significantly impede its use in China, where the government is committed to its use and has a range of domestic sources for the technology.
It is also to say that when new technologies have a wide range of possible uses, it can be difficult for states to draw bright lines—both domestically and internationally—about what uses are within bounds. Putting down markers about specific uses and contexts that are clearly out of bounds is a good place to start. Even if the U.S. government is imposing these sanctions for reasons that are not entirely salutary, sanctioning FRS providers in this case will provide a useful precedent for future cases.