I tried moving my legs and torso in the most contorted way possible. Every time, the software guessed that it was me. Your gait is like a fingerprint, unique and impossible to falsify.

I was being shown the new gait-recognition technology developed by Watrix, a Beijing startup founded by a computer scientist at Tsinghua and a few of his brightest students. Rather implausibly, one student explained that the technology could be used in a “smart” home to detect who has entered a room in order to adjust the temperature automatically. As I smiled, she admitted that the product’s real uses were for law enforcement and security.

As protesters in Hong Kong have demonstrated, facial recognition is not foolproof. You can use a mask or walk with your back to the cameras; even makeup can pose problems. Distance is a factor. At present, of course, Chinese facial-recognition technology has been rendered obsolete by the onset of the coronavirus epidemic. Public authorities have mandated the use of face masks. Suddenly, even the facial-recognition traffic lights installed in Shanghai to shame jaywalkers are no longer functioning. In Hong Kong, chief executive Carrie Lam, who pushed for a ban on face masks, is now wearing them.

Gait recognition does not suffer from any of these limitations. Simply put, it no longer requires any level of cooperation from those being surveilled. The computer scientist at Watrix quotes Shakespeare: “I know her by her gait.” Could there be a better proof that gait is superior to all alternatives?

Facial-recognition technologies have been receiving the brunt of the tech backlash. Many cities in the United States have banned them. The European Union recently admitted that a ban might become necessary, and news reports have shown how a smartphone photo taken on the street could be correlated with a trove of online images.

People might notice that they’re being watched by cameras, but they’re much less likely to be aware of sensors on the floor, another gait-recognition technology that Watrix is exploring. We’re fixated on facial recognition because it promises a measure of control. A face can look back and at least know when it is being watched. With gait recognition, that feeling of control is gone.

When I ask Qingqing, the computer scientist, how the software manages to identify a person by his or her gait, she says that there is no way to know. The algorithm analyzes the images it takes against the classifications it is given and then repeats the experiment until it starts to get it right—the proverbial black box, but rather literal in this case, as my gait was filmed and processed inside a dark, cramped room.

The team at Watrix is confident that gait recognition will survive unscathed even if facial recognition is banned in many parts of the world. It’s less intrusive, they say. On some level, that must be false, but it points to a deeper truth.

What gait recognition does is provide something that we are not very good at. We may recognize a loved one by the way that person walks, but our powers end there. The line in Shakespeare’s The Tempest would be absurd if it said that Juno was recognized by her face. But gait recognition is something unusual and novel.

If Watrix is correct, recognition technology based on deep learning will be all the more unstoppable, the less it resembles the normal toolbox of human skills. First, gait, then heartbeat patterns, and, eventually, microbiomes—every person emits about 36 million microbial cells per hour, and human microbiomes are unique—or odor biometrics.

What we fear is less the surrender of privacy than the rise of humanoid machines. We fear being usurped or replaced. That a human face could be transformed into data and fed into a machine algorithm is unnerving because it suggests that our deepest emotions and reactions are themselves a mechanical operation. To submit to the judgment of a machine on matters of sex and beauty, race and health, emotion and artifice, seems tantamount to abolishing humanity.

New technologies are just as intrusive as facial recognition, but they may superficially be less insulting to our individuality. Robots that look like us will always remain a threat. Machines whose abilities are genuinely unfathomable will be either welcomed or ignored. The artificial mind being created at dozens of startups like Watrix is something not just different from, but beyond, human experience.

Photo: Diy13/iStock


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next