Amanda Wagner JD Anticipated May 2023, Law & Public Policy Scholar
In November, Facebook announced that it was planning to shutter it’s eleven-year-old facial recognition system, deleting more than one billion users’ facial scan data in the process. Some may have believed the move signaled broader consensus regarding the potential privacy abuses, threats to civil liberties, and questionable accuracy of facial recognition technologies. After all, the announcement comes after a period of significant public discourse about the technology’s intrusive and inaccurate applications, with activists and police alike expressing fear regarding nonconsensual digital identification. The move by Facebook also follows in the footsteps of a slew of other tech giants. In June 2020, IBM announced that it would no longer offer facial recognition products, a decision which was quickly followed by statements from Amazon and Microsoft that they would place moratoriums on law enforcement access to their facial recognition programs. All three companies cited concerns regarding misuse as the reason for their policy changes.
However, the battle over the development and application of facial recognition technology is far from over. Less than one month after Facebook’s announcement, Politico reported that the facial recognition technology company Clearview AI had received a “notice of allowance” from the U.S. Patent and Trademark Office. As soon as the administrative fees are paid, Clearview AI will receive a patent which covers the company’s methods of providing information based on facial recognition. This patent will include the company’s highly controversial “automated web crawler” which scraps publicly available images from the internet to be analyzed as a part of the facial recognition process.
Clearview AI is likely the most notorious facial recognition technology company on the market. Founded in 2017, the company’s website touts its impressive search engine of over ten billion images “sourced from public-only web sources, including news media, mugshot websites, public social media, and other open websites.” According to documents obtained by Buzzfeed News, within Clearview AI’s first three years of operation it facilitated nearly 500,000 searches by 2,228 law enforcement agencies, companies, and institutions. Though the company no longer provides systems for private use, it continues to proudly advertise its relationship with law enforcement and allows potential law enforcement clients to request a thirty-day free trial through its website.
This is far from the first facial recognition patent to be granted – businesses across a variety of sectors have received about 5,000 facial recognition-related patents between 2015 and 2019. However, the Clearview AI patent appears to be the first one to legitimize the controversial method of utilizing public internet information for comparison on such a large scale. News of Clearview AI’s forthcoming patent should be a further incentive for policy-makers to take the necessary step to regulate the deployment of facial recognition technology.
While ad hoc and state-specific regulation has emerged and certain market actors have voluntarily removed themselves from the industry, the use and development of the facial recognition technology continues to proliferate. A June 2021 survey found that twenty federal agencies which deploy law enforcement were operating facial recognition technology systems. Of these agencies, thirteen reported use of non-federal facial recognition systems but were unable to provide information regarding which specific facial recognition technologies were being utilized by employees. Ten agencies reported use of Clearview AI, despite the fact that Clearview AI only recently subjected its algorithm to limited third-party testing. Additionally, ten federal agencies have already reported plans to expand the use of such technology through fiscal year 2023, with many planning to deploy new facial recognition systems.
Private companies are also deploying facial recognition technologies. Many large retailers have implemented facial recognition systems within their stores, intended to assist with theft prevention, targeted advertising, and employee monitoring. This widespread and unregulated use has resulted in a number of lawsuits by individuals falsely identified by facial recognition systems. Overwhelmingly, these bad matches have implicated persons of color for crimes they did not commit.
Given this reality, comprehensive regulation is necessary to ensure the continued growth of this type of technology occurs responsibly. In the absence of such oversight, unscrupulous actors have been allowed to profit and there has been little, if any, accountability for errors or harms stemming from the technology’s application. Though Clearview AI has suggested the company no longer has plans to develop its technology for commercial applications, there is little restricting such development. Comprehensive regulation, including limitations on permissible uses, must address these concerns while demanding improved reliability, transparency, and protections for individuals’ biometric data.
A Facebook whistleblower recently exposed the harms that may occur when a technology platform is allowed to operate without regulatory oversight. Despite significant data alerting Facebook to the consequences of their practices, operations benefiting the company’s shareholders and bottom-line took precedent. Without a national regulatory regime, facial recognition technology companies like Clearview AI are similarly beholden only to the self-interests of their shareholders. Since 2017, Clearview AI has grown and profited in the face of concerns regarding privacy violations, misuse, and inaccuracy due, in large part, to the use of controversial methods of data gathering legitimized by the forthcoming issuance of a patent. Congress must enact comprehensive legislation to prevent such companies from continuing to profit at the expense of both individuals and society.