As the European Union considers a five-year ban on facial recognition technology, affording lawmakers the opportunity to conduct research and craft meaningful regulation, revelations regarding the practices of an American facial recognition company should motivate Congress to consider its own temporary ban on the technology.
Facial recognition software has been controversial for some time now. The technology has the ability to compare photos taken by, say, a surveillance camera with a larger pool of photos — driver’s license pictures, mug shots, etc. The software then connects the images, thus giving the user the identity of the person captured on surveillance image.
Law enforcement and government officials have celebrated the technology’s ability to identify criminals, while civil liberty and privacy advocates fear facial recognition could infringe upon the rights of every citizen whose image is captured and examined.
To this point, facial recognition has relied upon preexisting photo databases, curated by law enforcement or the government. But Clearview AI, a relatively new facial recognition company founded by Australian entrepreneur Hoan Ton-That, has pushed the technology into disturbing new territory.
According to an investigative report published by the New York Times, Clearview’s software is groundbreaking in its scope. A user can upload a picture of a person, which is then cross-referenced with Clearview’s photo database — a database that has been constructed by scraping more than 3 billion images from social media sites like Facebook, YouTube, Venmo and more.
Clearview’s “innovation” has sparked a lot of interest in the U.S. law enforcement community. The company has partnered with more than 600 state and federal agencies throughout the country. Officials claim that Clearview has been used to solve crimes ranging from theft to fraud to murder. The Indiana State Police say Clearview helped solve a shooting within 20 minutes of the incident.
Despite the excitement the software has created among law enforcement, Clearview’s business model is rife with concerns related to civil liberties, criminal justice and privacy.
Even the most rigorously tested facial recognition systems are not 100% accurate. In fact, they are inaccurate an alarming amount of the time, particularly when identifying people of color. But Clearview’s system has not been scrutinized by independent third parties or industry experts. And because large databases increase the risk of misidentification, Clearview’s system could be even more likely to provide false positives, which could lead to wrongful arrests.
There are also many concerns about the data Clearview has accumulated, who may have access to it and how the software could be used. Any person who appears in a photo that has been posted on social media — in other words, most Americans — could be identified by Clearview. From there, the opportunities for abuse are myriad.
Congress should see the current conversation around Clearview and facial recognition as reason to take action. Clearview’s decision to transgress the unofficial rules of facial recognition have made it clear that real rules must be created, to protect citizens from unlawful intrusion by private companies, law enforcement and the government.
U.S. lawmakers should take a cue from the EU and impose their own temporary ban on facial recognition technology. The suspension would give researchers time to study facial recognition’s accuracy, its effects in other countries (facial recognition is already commonplace in China) and create recommendations for future regulations.
The dangers of facial recognition are too numerous to allow its spread to continue unabated. The U.S. is on a slippery slope to accepting a surveillance state. We need more research, evidence and debate before facial recognition can become the norm.