The Camera that Never Sleeps: Facial Recognition and the End of Australian Anonymity
Western Australia is preparing to introduce real-time facial recognition cameras in designated public areas; a move being described as an Australian first. Supporters present the technology as merely another tool in the fight against crime. We critics see something far more significant: a step-change in the relationship between citizens and the state. The issue is not simply whether the technology works, but what happens to a free society when every public movement can potentially be monitored, recorded, analysed, and archived?
For centuries, one of the unwritten liberties of public life was anonymity. A person could walk down a street, browse in a bookshop, attend a political meeting, visit a church, or meet a friend without creating a permanent government record of their whereabouts. Police could observe public behaviour, but they could not simultaneously identify and track thousands of people in real time. The practical limitations of surveillance acted as an important safeguard against abuse.
Facial recognition changes that equation. The technology transforms every camera into a potential identification checkpoint. Instead of observing behaviour, authorities can identify individuals automatically and continuously. The distinction is crucial. Traditional CCTV records events. Facial recognition identifies people.
Proponents often respond that law-abiding citizens have nothing to fear. Yet history suggests that powers granted for one purpose rarely remain confined to their original justification. Surveillance tools introduced to catch serious criminals frequently migrate towards broader applications. What begins as a system targeting violent offenders can gradually expand to encompass minor offences, regulatory breaches, political monitoring, or intelligence gathering. The temptation is obvious. Once the infrastructure exists, the pressure to use it grows.
Equally troubling is the issue of accuracy. Advocates sometimes speak as though facial recognition is infallible. It is not. Numerous studies have identified error rates that vary according to image quality, lighting conditions, database quality, age, ethnicity, and other factors. Even a seemingly tiny error rate becomes significant when applied to hundreds of thousands or millions of people. A system that is 99 percent accurate may sound impressive, but if it scans a million faces, thousands of innocent individuals may be incorrectly flagged.
The consequences of false identification are not trivial. A mistaken match can trigger police attention, questioning, searches, detention, reputational damage, or inclusion in intelligence databases. Unlike a simple clerical error, a facial recognition mistake carries the weight of technological authority. Many people assume that if the computer says it is you, then it must be true. Yet computers merely automate existing errors and sometimes create new ones.
There is also the problem of mission creep. Today's facial recognition systems identify individuals. Tomorrow's systems may classify behaviour, emotional states, social associations, or risk profiles. Artificial intelligence systems are increasingly capable of drawing inferences from vast quantities of data. A society that normalises mass biometric surveillance may eventually find itself living within a digital architecture of continuous assessment.
Supporters frequently invoke security. Security is important. Citizens have a legitimate interest in protection from violence and crime. But liberty is also important. A free society is not merely one in which crime is reduced; it is one in which citizens retain a sphere of private life free from constant observation. The challenge has always been balancing security and freedom. New technologies make that balance harder, not easier.
The deeper question concerns the presumption underlying the technology. Traditional policing begins with suspicion directed towards specific individuals. Mass facial recognition reverses that logic. Everyone becomes a potential suspect whose identity is verified continuously just in case they might be someone of interest. The burden of scrutiny shifts from the exceptional case to the entire population.
Perhaps the most important lesson from history is that surveillance powers should be judged not only by the intentions of current officials but by the possibilities they create for future governments. Technologies endure long after political leaders depart. A surveillance system established for noble purposes today may be inherited by less restrained authorities tomorrow.
Western Australia's experiment may therefore represent more than a policing initiative. It may mark a cultural turning point. Once anonymity in public spaces is surrendered, it is extraordinarily difficult to recover. The camera never sleeps, never forgets, and increasingly never fails to recognise. That future is incompatible with a genuinely free society.
