This post first appeared on Buzzfeed News.

Hong Kong’s High Court ruled Monday that a government ban on protesters wearing face masks was unconstitutional. The mask ban has become a flashpoint in the movement to defend democracy there, and is part of a growing battle between high-tech surveillance and evasion of such surveillance. Demonstrators have not only worn masks but also used umbrellas to shield themselves from surveillance cameras, destroyed CCTV cameras and “smart lamp posts,” and used lasers to neutralize them.

What many people do not realize is that a number of American states and cities also ban the wearing of face masks during protests. And in the United States, as in Hong Kong, the issue is likely to grow in significance due to the increasing prevalence of new surveillance technologies, especially surveillance cameras and face recognition.

Some of our anti-mask laws date back to the 19th century. A New York State law, for example, was originally enacted in 1845 to quash a sometimes violent rebellion by tenant farmers who dressed up as Native Americans to hide their faces. Some state laws apply narrowly (such as only during the commission of a crime) and others quite broadly. They typically include exceptions such as for holiday costumes, masquerade parties, sports and entertainment events, physical safety, protection from the cold, and religious beliefs.

Court challenges to American anti-mask laws have met with mixed success. A California law was struck down by state courts in 1979 in a challenge from Iranian-Americans who argued that it exposed them to retaliation by the Iranian government. Other courts have ruled that such laws are constitutional, most significantly the 2nd Circuit U.S. Court of Appeals, which upheld New York’s law in a case involving the Ku Klux Klan.

In fact, the KKK has been central to the history of mask laws in the United States, and that understandably complicates the issue for many people. A number of southern states, including Alabama, Florida, South Carolina, Virginia, and West Virginia, enacted mask bans in the late 1940s and early 1950s to repress the Klan. The political leaders in those states were acting not to help the Klan’s targets, but because they wanted to defend segregation as part of a “modern South” and felt that violent racist terrorism was making them look bad.

The political and technological contexts in which these laws are being applied are changing, however. Anti-mask laws today are being deployed more often against progressive protesters than Klansmen, and in ways that raise questions about selective prosecution.

New York State’s law was used to arrest Occupy Wall Street protesters, for example. In Alabama, police cited that state’s anti-KKK law to force people to remove their masks while protesting an April 2017 speech by a white supremacist. Alabama’s law was also used to arrest the organizer of a November 2018 protest over the fatal shooting of a Black man. In Virginia, several college students were arrested on felony charges and initially faced prison time for covering their faces while protesting a pro-Confederate rally. A legislator in Arizona proposed an anti-mask law, declaring that anti-Trump and Black Lives Matter protesters who damaged property were “absolutely” the “equivalent of the Ku Klux Klan.” (“Now, there are no hangings of white people, yet,” he added.)

At the ACLU, we continue to get queries from protesters and even from street performers dressed as superheroes who have been cited or threatened under a mask ordinance. But it’s the spread of facial recognition that is likely to raise the stakes around anti-mask laws the most. The more accurate and widespread the technology becomes, the more situations will arise where people won’t want to show their faces. The cameras that increasingly surround us will allow the police to cheaply and easily identify us — and who we’re with, even if part of a giant crowd.

It gets even worse when we realize that we can be identified that way repeatedly, with our presence recorded in databases and automated algorithms used to flag repeat protest activity, associational patterns, or anything else the authorities might want to know.

And of course it’s not just the authorities who can use face recognition. Even if we restrict its use by government, private parties who have access to photographs of a controversial event will be able to use them to identify us and harass us, retaliate against us at work, or worse.

If we allow government face recognition to continue to spread, the American context will increasingly resemble what we see in Hong Kong today: a struggle for power centered around who can observe and who can be observed. American laws should allow people the freedom to cover up their faces in protests or anywhere else. If necessary, rules should only target uses of masks that are intended to facilitate commission of a crime. We must maintain the freedom to publicly express ourselves while protecting our privacy, in a world increasingly designed to take it away from us.

“Look at the camera,” a Customs and Border Protection (CBP) officer told me as I approached the primary inspection point at the Paso del Norte port of entry.

“Is that the face recognition technology?” I asked. “If so, I want to opt-out.”

“Look at the camera.”

“I want to opt-out.”

“Look at the camera.”

On the evening of November 25, 2019, I crossed from Mexico into the United States. Signs in the port noted the new use of face recognition technology and United States citizens’ option to “undergo alternative procedures.” After handing over my U.S. passport card, and despite my repeated protests, the CBP officer took my picture anyway.

My data was likely processed through a system of databases, handled by the Department of Homeland Security (DHS), an agency with a long history of employing abusive surveillance techniques. In the last year, DHS has unlawfully tracked journalists and advocates, and was the subject of data breaches that exposed the private information, including photos, of thousands of travelers.  

CBP, a component agency of DHS, began piloting the use of “biometric facial comparison,” or face recognition technology, in September 2018 at pedestrian crossing lanes in San Ysidro, California and vehicle lanes in Texas. On November 22, 2019, CBP announced the expansion of the technology to additional pedestrian lanes at ports in Texas, including the Paso del Norte port of entry in El Paso. This expansion came despite an internal review raising concern about the accuracy of such technology and despite serious problems identified by privacy experts about prior expansions at airports.  The agency intends to expand the use of face recognition nationally at airports — despite the documented concerns.

CBP deploys a variety of surveillance technologies at the border, claiming national security justifications, but Congress has not explicitly authorized the use of face recognition technology in the immigration context. Congress has mandated that the DHS collect biometric information to track travelers entering and exiting the United States to identify those who overstay their visa, but fingerprints — and other less troubling methods — could achieve compliance without the worries surrounding face recognition.    

Face recognition is one of the most dangerous forms of biometric tracking and carries a greater potential for growth into a widespread tool for spying on people as they move. Face recognition can be used for surveillance through public video cameras — mapping a person’s movement without their knowledge or consent and raising serious Constitutional concerns.

Photos collected by state motor vehicle agencies provide another source of data that could easily be coupled with face recognition to create a comprehensive surveillance system equipped to track U.S. citizens and immigrants alike throughout the country.

Given the many concerns and shortage of mechanisms to safeguard against abuse, immigration agencies should suspend their use of the technology at ports of entry. 

CBP claims the technology will facilitate faster border crossings but the technology is inaccurate, exposing crossers to further inspection if the system falters. Studies also suggest the technology is racially biased, with error rates rising significantly when applied to people of color.  

If I, carrying all the privilege of a white American lawyer, could not opt-out of the invasive technology, what chance do other travelers — and particularly people of color — have to assert their rights before an agency patterned on racial profiling and harassment? Indeed, many other travelers have been forced to submit to invasive face recognition — despite the agency’s promises that anyone can opt-out.

CBP clearly and consistently states that “it is not mandatory for U.S. citizens to have their photo taken” and if they wish to opt-out they should “advise the CBP officer when they approach the primary inspection area.” While no one should be subjected to this technology, CBP must minimally provide a meaningful opt-out option that does not mean hours of delay, and must train its agents on this policy. Even the best policies are meaningless if government agents are free to disregard them with impunity.

The CBP officer I encountered last week ignored my repeated protests, claimed ignorance of the signs plastering the port, and told me I could not opt-out. “Why are you so concerned? We have all your information anyway,” was the last thing the officer said before waving me through.

My concern is one we should all share: The continued expansion of surveillance technology at the border, under the guise of efficiency and security, signals the erosion of our privacy rights and the building of a system of government surveillance capable of intrusion in our everyday lives. Taking away every meaningful option to avoid new forms of surveillance simply cannot be an accepted border reality. The Constitution protects us all, even at the border.