You are being watched. That is the baseline condition of modern public life. The camera above the shop door, the transport authority’s dome on the station platform, the neighbour’s video doorbell, the police vehicle’s automatic scanner—they all record. The common assumption is that privacy in public is a lost cause, a relic of a less connected past. The real risk, however, is subtler. It is not the isolated recording but the system that stitches those moments together, transforming scattered observations into a searchable, persistent biography of your movements and associations. Your goal shifts from impossible invisibility to a practical one: making that joining operation harder, slower, and subject to clear rules.
The network is the surveillance
To defend yourself, you must first understand what you are defending against. The individual camera is a sensor; the network of databases and matching algorithms is the true engine of surveillance. Each type of camera captures a different slice of your identity.
Fixed commercial cameras, like those in shops or on office buildings, create a localised log of presence. Their primary purpose is loss prevention or traffic monitoring, but the footage is often retained and can be accessed later. Transport network cameras, covering buses, trains, and stations, track movement through specific infrastructure. They are frequently used for operational safety but also feed into wider security monitoring.
Residential cameras, especially internet-connected video doorbells, have moved observation into residential streets. They capture not just visitors but passers-by, often uploading footage to cloud servers with retention policies set by the manufacturer and the homeowner. Police body-worn cameras and vehicle systems add an official layer, while automatic number plate reader (ANPR) systems, as detailed in my earlier piece on licence plate readers explained, create a precise, timestamped map of vehicle movements, often stored in searchable databases for extended periods.
The critical point is that in isolation, a clip from any one of these sources may be meaningless. You are just a figure in a crowd. The threat emerges when these disparate data points can be reliably joined. That joining turns “a person” into you.
The joining mechanisms: face, phone, and plate
Three primary keys facilitate this joining operation: your face, your mobile phone, and your vehicle registration plate. Alone, each is a powerful identifier. Combined, they render anonymity in public spaces a fiction.
Face recognition is the most direct link. A clear capture of your face from a shop camera can be matched against a reference image—perhaps from a driving licence database, a social media profile, or a previous police encounter. This process, which I’ve described as the face becoming a password in the face is the password now, allows an operator to put a name to the figure in the footage instantly. It bridges the gap between anonymous observation and positive identification.
Your mobile phone is a persistent, noisy beacon. Even if you are not making a call, it constantly signals its presence to nearby cell towers and seeks out known Wi-Fi networks. This generates rough but continuous location data for your network provider. In more targeted scenarios, law enforcement or other actors with specialised equipment can use International Mobile Subscriber Identity (IMSI) catchers to harvest unique phone identifiers from a crowd, placing a specific device—and likely its owner—at a precise place and time. This data can then be correlated with camera footage.
Your vehicle’s number plate is a unique, openly displayed identifier. ANPR systems log the plate, time, and location. This creates a highly reliable trail. If your face is seen near a car that is later logged by an ANPR camera, the two records can be linked. Similarly, if your phone’s location data places you moving at vehicle speed along a route where your car’s plate was read, the connection is strongly implied. The cost of this constant findability, both socially and personally, is explored in the cost of being findable.
What you can reasonably do
You cannot walk around with your face obscured, your phone off, and without using a car without attracting more attention than you deflect. Practical defence, therefore, focuses on reducing the low-hanging fruit, making casual joining harder, and limiting the data you voluntarily contribute to these systems.
Manage your phone’s wireless signals. Disable Wi-Fi and Bluetooth when you are not actively using them. Your phone will not then probe for every remembered network or visible device, creating fewer data points that can be logged. Review and periodically clear your device’s location history if your operating system stores one. Consider using a faraday bag or sleeve if you need absolute certainty your phone is not transmitting, but recognise this is a specialist measure for specific situations, not daily life.
Be conscious of payment methods. Using a contactless card for every small purchase creates a detailed financial diary that can be correlated with camera locations. Using cash for routine, low-value transactions breaks that chain. It is not about hiding illicit activity but about not creating a perfect, bank-verified record of your daily movements.
Think before you post. That photograph shared on social media, geotagged to a café at a specific time, is a gift to any profile-building system. It confirms your presence, your appearance, and your associations at a known moment. The guidance in before you post that photograph applies doubly here: you are not just sharing with friends but potentially with every entity that scrapes public data.
Know and use your existing rights. In jurisdictions with data protection law, you have a right to access personal data held about you. This can include requesting footage from a private business’s camera system that shows you. Making such requests, while often laborious, reinforces the principle that this data is subject to rules. It also gives you visibility into what is being stored.
What not to rely on
Certain defences are either ineffective or carry disproportionate risk. Do not rely on simple visual disguises like hats or sunglasses against modern face recognition. These systems can often infer identity from partial features, gait analysis, or contextual data. The effort and social cost rarely match the minimal protection gained.
Do not assume that because you are in public, you have no privacy rights. The legal concept often differs from the technical reality. While you may have a reduced expectation of privacy regarding being observed, you retain rights regarding how that recorded data is processed, stored, and combined.
Most importantly, do not conflate evasion of commercial or general surveillance with evasion of lawful police investigation. The techniques discussed here are for reducing your exposure to pervasive, non-specific data harvesting. Actively obstructing a police investigation is a separate matter with serious legal consequences. The goal is privacy, not impunity.
Rules are the real defence
Your individual actions can raise the cost of profiling you, but the most significant protections are collective and legal. These are the rules and norms worth understanding, advocating for, and demanding from those who operate surveillance systems.
Short, mandatory retention periods for non-essential footage are fundamental. There is no legitimate need for a supermarket to keep general entrance footage for more than a few days unless an incident is flagged. Legislation should mandate automatic deletion after a brief period, drastically reducing the pool of historical data available for retrospective mining.
There must be clear, legislated limits on the use of live and retrospective face recognition. Its use by state authorities should require judicial authorisation for specific, serious investigations, not be a tool for general population monitoring. Its use by private entities in public spaces should be heavily restricted or banned.
Access to surveillance databases must be logged and auditable. Every query run against an ANPR database or a police camera archive should leave a trace—who accessed it, when, and under what justification. This creates accountability and deters casual fishing expeditions.
Finally, there must be transparency about what systems are in use. The public has a right to know if a street is covered by cameras equipped with real-time analytics or if a transport network is conducting passive face recognition trials. You cannot consent to or contest a system whose existence is hidden.
Questions people ask
Can I legally film back, or film the cameras?
In most public places, you can generally photograph or film anything in plain view, including surveillance cameras on public buildings. Filming back at a police officer or inside a private building like a bank may have restrictions. The legality often hinges on whether you are obstructing duties or violating specific local laws. Your right to document your surroundings is a powerful counterbalance, but it is not absolute.
Do face masks or makeup fool recognition systems?
Modern systems are highly resilient to simple obstructions. While a medical face mask might reduce accuracy, it is not a reliable defence. Specialised makeup or patterns designed to confuse algorithms can work against some versions, but these systems evolve rapidly. Relying on appearance alteration alone is a fragile strategy compared to advocating for legal restrictions on the technology's use.
What should I ask my local council about surveillance?
You should ask for a public register of surveillance camera locations, especially those owned or funded by the council. Inquire about the stated purpose for each system, its data retention period, whether it uses any biometric processing like face recognition, and the audit process for access logs. Demand clarity on who has access to the feeds and under what authority.
Are there apps that warn me about nearby cameras?
Some community-driven mapping applications attempt to log camera locations, but their coverage is incomplete and unofficial. Relying on them creates a false sense of security. A more effective approach is to operate on the general assumption that in urban centres, your entry into most commercial zones and major streets will be recorded.
Close
Privacy in public is not about disappearing. It is about maintaining the healthy separation between the different roles you play in society—shopper, commuter, protester, pedestrian—and preventing the constant, automated fusion of those roles into a single, permanent dossier. The technical capability to build that dossier now exists, woven from camera feeds, plate reads, and phone signals.
Your practical daily habits can fray the edges of that net, making you a slightly harder target for casual profiling. But the strongest thread you have is the demand for governance. The fight for privacy on the surveilled street is won less by wearing a larger hat and more by insisting on short retention periods, banned uses for face matching, and transparent logs. It is the difference between being seen, which is inevitable, and being permanently filed, which is a choice a society makes. Your task is to argue, persistently, for the rules that make that filing harder to justify and more expensive to perform.