While recent developments may suggest crackdowns on certain uses of facial recognition systems, these systems and the underlying technology are here to stay. Regulatory proposals governing facial recognition systems are forthcoming, and the issues surrounding this technology will remain relevant and widely debated.
What is a facial recognition system?
Facial recognition systems are AI-based systems configured to match a person's face in a digital image with stored image(s). Facial recognition systems are commonly used in smartphones, software applications, automated technologies such as robotics, and the like.
How does a facial recognition system work?
A facial recognition system performs AI-based pattern recognition of human faces using a common machine learning technique known as artificial neural network. The initial step is Face Detection, where a portion of the image containing the face is separated from the background of the image to determine the exact location of the face on the image. In the next step called Feature Extraction, the portion of the image containing the face is aligned to account for face pose, face size, and image properties to accurately localize and extract facial features (e.g., eyes, nose, mouth, geometric shapes, and distances) through an artificial neural network. The final step is Face Matching, where the output of the artificial neural network is digitized image data, which is then matched against a stored image or a database of images.
What happened recently?
On November 2, 2021, global social media platform, Facebook (n/k/a Meta), announced that it would end use of its facial recognition system and delete more than a billion users' facial recognition templates that it stores to match faces with photos and videos. Facebook cited privacy and regulatory challenges, including ongoing government investigations and a class-action lawsuit, as part of its decision to end use of facial recognition systems.
On November 3, 2021, the Australian government ordered international facial recognition company Clearview AI to cease collecting images in Australia and destroy collected images because Clearview had breached Australians' privacy by collecting and sharing through its platform face-identification information without consent and by unfair means. Clearview AI holds a database of more than three billion images collected from the internet.
Why is this important?
While technology of facial recognition systems has developed rapidly, so have the legal issues surrounding its use. Companies using facial recognition systems have faced legal challenges and wider criticism in society. Privacy remains the foremost legal issue, and it stands on unclear grounds. There remains no US federal law regulating the use of facial recognition systems, and only a few states such as Illinois, have a biometric privacy law at the state level.
Further, police use of facial recognition systems has revealed racial bias inherent in these systems, which implicate civil rights issues. In 2020, technology companies like Amazon, Microsoft, and IBM decided to cease selling their facial recognition systems to law enforcement agencies citing concerns about false identifications and inaccurate racial profiling.
What next?
The technology and applications of facial recognition systems are likely to continue growing. Facebook's (n/k/a Meta) latest decision is limited to its social media platform, and the company will likely continue using DeepFace, the complex algorithm powering its facial recognition system, for other purposes like its Metaverse project. On the legal front, we expect to see regulatory proposals addressing issues raised by facial recognition systems both at the federal and the state level. These proposals may eventually help develop a common consensus on how to govern the use of facial recognition systems, especially as their use will continue to remain prevalent.