Technology vs. Crime: Is Innovation Making the World Safer?

For centuries, crime prevention depended largely on physical presence: an officer on the beat, a guard at the door, cash locked in a safe, witnesses interviewed after the fact, fingerprints compared by hand. Today, a person committing a crime may be captured on several cameras before reaching the scene, their vehicle identified automatically, their phone generating a location trail, and their financial transactions leaving a digital record. The instinct behind this article's starting question — that technology has made crime harder to get away with — is a reasonable one, and a large body of criminological evidence supports it.

But the strongest version of that claim is not "technology alone caused the global reduction in crime." Crime rates move with policing strategy, demographics, economic conditions, urban design, drug markets, incarceration policy and legislation, among other variables — no single factor, technology included, explains the whole trend. The version of the thesis that actually holds up under scrutiny is narrower and, in some ways, more interesting:

This matters for how a business, a government agency, or a security planner should read the statistics that follow. The decline in burglary or car theft is real and well evidenced. So is the rise in cybercrime. Reading only one half of that story — "crime is disappearing" or "crime is exploding" — leads to the wrong conclusions. Reading both halves together leads to the right question for 2026 and beyond: not whether to invest in security technology, but which layer of the technology stack — physical, digital, or both — a given organization is actually exposed on.

England and Wales offer one of the world's clearest, longest-running datasets on this question, because the Crime Survey for England and Wales (CSEW) has tracked household and personal crime experience — independent of what gets reported to police — since 1981. The picture it shows is a dramatic, three-decade decline in traditional property crime.

It would be wrong to attribute this entire decline to technology alone — the same period saw major shifts in policing strategy, demographics, drug markets and incarceration policy. But an influential and still-active strand of criminological research, sometimes called the Security Hypothesis, argues that improved security — much of it technological — played a materially important role in reducing the opportunity to commit common property crimes. Researchers studying household burglary, for instance, have linked improved combinations of door locks, window locks and security lighting to falling burglary rates, while the vehicle-crime literature is even more specific: a systematic review of 16 international studies across the UK, Germany, Australia and the USA found that 15 of the 16 showed electronic immobilisation successfully reduced vehicle theft, with the effect strongest for temporary ("joyriding") theft and somewhat smaller — but still present — for permanent theft.

The lesson generalizes well beyond cars: when technology increases the difficulty, risk and cost of committing an offence, some crimes become measurably less attractive to opportunistic offenders. That single mechanism recurs throughout the rest of this article, in CCTV, cashless payments, biometrics and AI alike.

Perhaps no technology has changed everyday criminal behaviour as visibly as CCTV. A potential offender entering a modern shopping centre, hotel, metro station, airport, residential tower or city centre must increasingly assume that someone — or something — is watching. Today's surveillance ecosystem typically combines high-resolution cameras, automatic number-plate recognition, AI-based object detection, behavioural analytics, central command centres and cross-camera tracking, not the isolated, low-resolution cameras of a generation ago.

CCTV has two distinct crime-control functions that are worth separating clearly:

• Deterrence — the possibility of being recorded raises the offender's perceived risk before an offence takes place, shifting the classic "is there an officer nearby?" calculation toward "how many digital traces will I leave behind?"
• Investigation — even where deterrence fails, recorded footage helps reconstruct what happened: a camera identifies a face, another the vehicle, a traffic camera records its direction, plate recognition confirms the registration, access records establish entry time. Individually modest, these fragments combine into a digital reconstruction of the crime.

The smartphone may be one of the most consequential forensic devices ever created, not because it was designed for investigation, but because it continuously interacts with the infrastructure around it. Depending on the circumstances and applicable legal process, evidence available from a device can include cell-tower connections, GPS data, Wi-Fi and Bluetooth interactions, application records, communication metadata, photographs, cloud backups and transaction records. A suspect may believe there were no witnesses; their own device may have quietly created a timeline instead. Investigations that once relied almost exclusively on eyewitness testimony can now combine physical evidence with this kind of digital corroboration.

GPS and connected telematics extend the same logic to physical assets. Vehicles, commercial fleets, shipping containers and consumer electronics can increasingly be located digitally — fleet-management systems record location, route, speed, ignition status and geofencing violations, while a stolen phone or tracked vehicle can, in effect, reveal where it has been taken. This does not make theft impossible, but it materially increases the probability of recovery and raises the risk profile of stealing anything that can phone home.

Of all the evidence connecting technology directly to falling crime, vehicle security produces the clearest single data point. Older vehicles could often be started using relatively simple mechanical techniques; modern vehicles introduced electronic immobilisers, transponder keys, alarm systems, central locking, GPS tracking and, more recently, connected vehicle services and remote lock-down. UK Home Office research estimated that electronic immobilisers alone accounted for roughly 25% to 50% of the reduction in stolen vehicles up to 2013, and — as noted above — a review of 16 international studies found 15 confirmed the effect.

The story has a second chapter worth flagging for a business or fleet-security audience: keyless entry systems introduced a new vulnerability — "relay theft," where a car's wireless signal is intercepted and amplified — that has eroded some of the gains made since the 1990s and helped drive newer UK legislation banning the possession and sale of electronic vehicle-theft devices. The takeaway for any organization deploying connected hardware is general, not specific to cars: security technology is not a one-time fix; each generation of protection tends to provoke a corresponding generation of workaround, which is why layered defenses — mechanical, electronic and behavioural — consistently outperform any single control.

For most of history, robbery had an obvious economic logic: people carried cash, retailers accumulated it through the day, and cash is immediately usable, portable and hard to trace. Digital payments changed that environment by letting large amounts of wealth move without physical currency changing hands at all. There is credible empirical evidence for the resulting effect on street crime, not just intuition.

The mechanism generalizes: credit cards, bank transfers, mobile wallets and instant payments all reduce the amount of physically stealable money circulating through daily life. That is an unambiguous societal benefit — but, as the next sections make clear, it does not remove the incentive to steal money. It relocates the target from a physical wallet to a digital account.

Fingerprints have supported criminal investigation for over a century; what technology transformed was the scale and speed of comparison. Digitised fingerprint databases let investigators search enormous record collections far faster than manual methods ever allowed, and modern biometric systems extend that logic to facial recognition, iris recognition, voice characteristics and behavioural biometrics. A suspect can change clothing, use an alias, dispose of a vehicle or cross a border — but biometric characteristics are considerably harder to alter, which is why, when integrated responsibly with national and international databases, biometrics can dramatically accelerate identification in cases that would otherwise stall.

Modern life continuously produces data, and a single offence can now generate a chain of independent, corroborating records — none of them individually decisive, but powerful in combination.

This represents perhaps the biggest shift in criminal investigation in modern history. The operative question is moving from "Was anybody watching?" to "Which systems recorded what happened?" — and the answer, increasingly, is several at once.

The UAE offers a particularly concrete, current example of technology-enabled public safety. Dubai has invested heavily in smart-city infrastructure, integrated surveillance and command-and-control systems. Dubai Police's Oyoon ("Eyes") initiative is an integrated surveillance and security ecosystem involving government, semi-government and private-sector partners, designed to use advanced technologies and AI to prevent crime, improve security and speed up incident response. Dubai Police has also publicly discussed using AI-enabled CCTV for capabilities including identifying wanted individuals through facial recognition and behavioural analysis. The real transformation is not the raw camera count — it is integration: a standalone camera only records, while a connected security platform understands relationships between events across a city.

The significance of the case extends beyond the value of the goods recovered. It demonstrates a structural change in what policing can achieve when surveillance, analytics and case coordination are integrated: investigations that historically might have taken days, weeks or months can, in the right circumstances, compress into hours. Technology in this model does not replace investigators — it dramatically expands what they can see and how quickly they can connect it.

Traditional CCTV has one fundamental limitation: someone has to watch it. AI changes the economics of surveillance by letting computer-vision systems continuously scan for unusual movement, abandoned objects, wanted vehicles, restricted-area entry or crowd anomalies, and surface only the events that need human attention. That shift — from asking humans to watch every feed, to asking AI to flag what matters — is what turns CCTV from passive recording into proactive detection infrastructure.

Data analytics adds a second capability: identifying patterns across historical incidents — where burglaries are increasing, at what times, by which methods, whether particular vehicles or networks recur — so patrol resources can be positioned more effectively. The realistic objective of predictive policing is not to name who will commit a crime, but to identify risk patterns that help allocate limited resources sensibly.

Crime-prevention technology has moved inside the home as well as the street. Smart doorbells, video intercoms, motion sensors, smart locks and cloud-recorded video mean an attempted residential burglary can generate evidence — a face on a doorbell camera, a vehicle on a neighbour's camera, a timestamp from a motion sensor preserved in the cloud even if the local device is destroyed — before the offender ever gets inside. Technology here changes both deterrence and evidence preservation simultaneously.

Banks show the same pattern from a different angle. Historically, banks represented concentrations of physical cash; modern banking increasingly represents data and account balances instead. Physical security technology, cash-management systems, CCTV and digital transaction records have all changed the traditional armed bank-robbery environment — but criminals adapted rather than disappeared. Instead of walking in with a weapon, an attacker is now more likely to target a bank through phishing, malware, stolen credentials, social engineering, business email compromise or SIM swapping. The target — money — has stayed the same. The attack surface has moved.

Here is the paradox this article's whole argument turns on: technology prevents crime, and technology also creates new crime. As physical money becomes digital, criminals attack digital accounts. As identity becomes digital, criminals steal identities. As businesses move online, criminals attack networks and supply chains. As people communicate through social media, criminals use social engineering. As artificial intelligence becomes more powerful and accessible, criminals use AI too. Crime, in other words, has not simply declined. It has migrated — and the scale of that migration is now large enough to show up clearly in national statistics.

The scale of this migration is now well documented by two of the most credible sources available: the FBI's Internet Crime Complaint Center (IC3) in the United States, and INTERPOL at an international level.

The three most commonly reported categories to IC3 in 2024 were phishing/spoofing, extortion and personal data breaches. INTERPOL's 2025/2026 Asia and South Pacific Cyberthreat Assessment, covering January 2024 to March 2025 across 18 member countries, adds further texture: the region recorded more than 135,000 ransomware-related attacks in 2024, a 92% surge in DDoS attacks, and a 600% increase in deepfake-related discussion on cybercriminal forums between February and June 2024. Cybercriminals are increasingly combining artificial intelligence, deepfakes, automated phishing, ransomware-as-a-service, synthetic identities and stolen databases into scaled, industrialized operations — the criminal, in short, is becoming technological too.

Artificial intelligence is increasingly operating on both sides of the law simultaneously, which is why the next phase of security is likely to look less like "humans versus criminals" and more like "defensive AI versus criminal AI."

The realistic future of security, on this evidence, looks like AI defending society against AI-enabled crime — which is exactly why the governance question raised earlier in this article, and revisited below, is not optional.

One useful way to synthesize everything above is a simple conceptual equation borrowed from rational-choice criminology: an offender's motivation is roughly a function of the expected reward, divided by the risk of getting caught, adjusted for how accessible the target is. Technology has moved all three variables — mostly against the offender, for the crimes covered in this article's first half.

For the crime categories where all three variables have moved this way — car theft, burglary, street robbery of cash — the rational attractiveness of the offence has genuinely fallen, and the statistics reflect that. For crimes where reward and accessibility are still favourable to the offender — scams targeting digital accounts, synthetic identity fraud, business email compromise — the same equation explains why criminal activity has concentrated there instead.

A technologically secure society must not automatically become a surveillance society. Facial recognition, biometrics, location tracking and AI analytics raise legitimate, recurring questions around privacy, consent, data retention, algorithmic bias, false identification, cybersecurity of the surveillance systems themselves, and accountability for how the data is used and by whom. The objective responsible governments and organizations pursue is not maximum data collection — it is intelligent, proportionate and accountable security. The safest smart city is not necessarily the one collecting the most data; it is the one using the appropriate data responsibly, securely and effectively, with human officers remaining responsible for decisions, intervention and due process at every stage.

The evidence in this article supports a clear, if carefully bounded, conclusion: technology has contributed materially to making several major categories of traditional crime more difficult, more detectable and less economically attractive. The decline in vehicle theft tied to electronic immobilisers is the strongest single data point; the long-term reduction in burglary and property crime alongside improved household security is another; and research connecting reduced cash circulation to reduced street crime provides a third, independent line of evidence. CCTV, GPS, digital payments, smartphones, biometrics, connected databases, smart buildings and AI continue that same trajectory, and Dubai's Oyoon programme and its rapid resolution of the AED 12 million oud case illustrate what integrated, well-governed technology can achieve operationally, in hours rather than weeks.

At the same time, the criminal economy has adapted rather than retreated. The thief who once stole a wallet may now target a password. The fraudster who once forged a cheque may now generate a deepfake. The gang that once targeted a physical bank branch may now target its digital infrastructure from another continent entirely — the FBI and INTERPOL data on cybercrime losses make that migration impossible to ignore.

For businesses, government entities and individuals in the UAE, the practical implication is straightforward: physical-security technology and cybersecurity are no longer two separate budget lines to be prioritized against each other — they are two halves of a single, connected risk. An organization that hardens its offices with access control and cameras while leaving its accounts, identities and networks unprotected has not reduced its risk; it has simply relocated it. The next question is not whether technology can make an organization safer, but where the current gap between the physical and digital layers actually sits — and that is precisely the kind of assessment Professionals Lobby's AI, IT and legal advisory teams are built to run.