By: Selena Liu

History of Facial Recognition Technology: The Very Beginning

Facial recognition technology has been in development since the 1960s, and initially, this technology was very manually intensive. In the 1990s, facial recognition underwent a significant revolution and expanded to data sets such as the Face Recognition Technology (FERET) database. FERET was completed in 1998 after the U.S. Department of Defense invested over $6.5 million. The database consisted of 14,126 facial images of 1199 individuals captured between 1993 and 1996 with their consent.

After FERET’s launch, many more datasets and evaluations, such as Face Recognition Vendor Tests, were developed to provide law enforcement agencies and the U.S. government with the information necessary to utilize facial recognition technology effectively. However, datasets such as Labeled Faces in the Wild began collecting facial photos from the internet in 2007, bringing into question issues of privacy and consent. Subsequently, databases and datasets, such as Clearview AI, were created for law enforcement with over 3 billion images to help solve cases involving shoplifting, identity theft, credit card fraud, murder, and child sexual exploitation. 

Potential Downsides to Facial Recognition Technology and Policy Reactions

There are potential downsides associated with the increased utilization of facial recognition technology by law enforcement. These include serious privacy invasions, and differential error rates by race and gender. The latter is particularly harmful, as certain groups are more likely to be falsely accused or criminally charged.

Recognizing the potential for vast privacy abuses, San Francisco’s legislature voted to ban city agencies from using the technology in 2019. Around two dozen cities in the U.S. have followed suit. Federally, bills in Congress were introduced in 2022 and 2023 to limit law enforcement use of facial recognition technologies, impose restrictions on federal, state, and local government entities, and ban the TSA from using these technologies.

New Frontier in Facial Recognition Technology: Public Access 

However, there is a new development regarding who gets access to facial recognition technology. Publicly accessible websites, such as PimEyes and Facecheck.id allow any member of the public to upload an image of someone. For a fee, anyone can access a grid of photos that are deemed to contain faces similar to the image that they upload, with links to where those images appear on the internet. Although the owner of PimEyes claims that users are supposed to only search for their faces, or the faces of those who have consented, there are no controls in place to prevent people from searching for the faces of strangers, or those of people who have not consented. 

Potential Regulations to Public Access to Facial Recognition Technology and Why Does it Matter?

Advances in facial recognition technology have outpaced laws governing its use, as bans, limitations, and proposed legislation found online are targeted toward government entities and law enforcement. Although there is progress in limiting its usage in various states, none of the current state laws address members of the public using websites such as PimEyes or Facecheck.id to identify other members of the public using merely a photo. Recently, a class action lawsuit was filed in Illinois, alleging that several facial recognition companies, including Pimeyes, violated the Illinois Biometric Information Privacy Act (BIPA). They specifically allege violations of BIPA’s requirements for informed written consent before collecting data such as facial photos. This could be a possible legal blueprint to require all publicly available facial recognition websites to obtain informed written consent from individuals before those websites can use their internet photos. 

Though publicly available facial recognition websites do not include photos from social media websites such as Facebook, the lack of restrictions against these facial recognition websites could lead to severe ramifications. Any individual could potentially find out who someone is by merely taking a photo of their face while walking on the street. It is increasingly likely in this day and age that the average person has at least one picture of themselves on the internet through being at a public event or having a professional headshot published on their workplace website. As of December 2022, only 14 percent of adults in the U.S. stated that there are no photos of themselves online. 

As facial recognition technology continues to progress, the amount of photographs in datasets rises. This allows many more members of the public to access data that could help them with stalking, doxing, and finding explicit photos of minors. Despite claims from PimEyes and Facecheck.id that they allow users to request removal of their photos, such removal requests do not extend to the third-party websites hosting those photos. Even if removal requests are complied with, it is unclear how many people are aware of this option as it is not widely advertised. Therefore, it could be almost impossible for anyone to fully control where their images can be used or prevent stalking.  

Conclusion

Facial recognition technology is a powerful tool that could be used by law enforcement and government entities to solve cases. However, it is also ripe for privacy abuse, consent issues, and misidentifications. As facial recognition technology becomes publicly available, a whole host of additional problems, such as increased ease of stalking and doxing, are becoming more prevalent. However, the law has not been keeping up with addressing these issues. 

To combat the problems around public access to facial recognition, states and the federal government must develop laws similar to Illinois’ BIPA and robust enforcement mechanisms. Websites must be deterred from becoming so accessible to the general public or using photos without informed written consent. Facial recognition technology will inevitably continue to revolutionize. Minimizing public access by addressing the legal gaps in laws regulating this technology is one of many keys to its continued potential as a public safety tool. 

#WJLTA #privacy #facialrecognition #technology