The ability to identify a person from their face in real time, at scale, in public spaces, has moved from science fiction to routine deployment. It is used by police departments, airports, concert venues, retail stores, and schools. The technology works — imperfectly, unevenly, and with consequences that are only beginning to be understood.
The book explored surveillance through Minority Report and Ghost in the Shell, and the Ubiquitous Surveillance page covered the infrastructure of constant monitoring. Since 2018, facial recognition has gone from a capability that existed to a capability that is deployed everywhere.
Police departments across the US, UK, and Europe use facial recognition to identify suspects from surveillance footage and in real-time at public events. Airports increasingly use it for boarding and border control. Retailers use it to identify known shoplifters. China has deployed facial recognition at a scale unmatched elsewhere, integrated into its broader surveillance and social credit infrastructure.
Clearview AI became a lightning rod for the field when it was revealed to have scraped billions of photos from social media — without consent — to build a facial recognition database used by law enforcement. The company's approach crystallized the tension: the technology is powerful and useful for identifying criminals, but the data practices that make it possible are invasive and ethically troubling.
Accuracy disparities are well-documented. Multiple studies, including by the National Institute of Standards and Technology (NIST), have found that facial recognition systems perform significantly worse on darker-skinned faces and on women. This means the communities already subject to disproportionate surveillance are also subject to disproportionate misidentification — a compounding of existing injustice.
The regulatory response is fragmented. Several US cities, including San Francisco and Boston, have banned government use of facial recognition. The EU AI Act classifies real-time biometric identification in public spaces as high-risk and restricts its use. Other jurisdictions have imposed no restrictions at all.
The book's Surveillance, Privacy, and Control framework identifies the core tension: surveillance technologies can enhance security and enable beneficial services, but they also erode the space for anonymity, dissent, and private life. Facial recognition intensifies this tension because it operates passively — you do not have to do anything, carry anything, or consent to anything. Your face is your identifier, and it is always visible.
The normalization dynamic the book describes — how societies gradually accommodate technologies that would once have been unthinkable — is playing out in real time. Facial recognition at airports felt intrusive when it was introduced. It is now accepted by most travelers without thought. Each expansion of biometric surveillance becomes the new baseline from which the next expansion is measured.
Everyone Has a Role — the book's insistence that technology governance is not just for experts — is directly relevant. Decisions about where and how facial recognition is deployed affect every person who walks through a public space, which is to say, everyone. See Why does it feel like nobody asked me about any of this?