overview

law

issues

Australia

studies
















related
Guides:

Privacy

Security
& InfoCrime






related
Profiles
& Notes:

Australian
Privacy
regimes

Surveillance

Travel

Assembly

ID Crime

Data
trading

     issues

This page considers automatic number plate recognition (ANPR) system issues.

It covers -

• introduction - making sense of ANPR policy questions
• issues - four key issues
• fuzziness - ANPR does not necessarily provide an authoritative match between a driver and the vehicle owner
• regulatory creep - discovering new uses for ANPR outside initial regulatory frameworks
• aggregation - centralised data collection and clearinghouses
• marketing - targeting consumers
• privacy - questions about data integration and trading

     introduction 

In essence, ANPR is a tool, rather than an outcome or a solution. The way is which the tool is used - or perceived to be misused - determines whether or not it is benign. Concerns regarding abuse relate to operational protocols for deployment of the cameras and frameworks, particularly statutory restrictions, on the mining and exchange of data gained from the cameras.

A key assertion by proponents of ANPR is that the technology greatly facilitates "intelligence-led policing", a term that is attractive (given positive associations in advanced economies regarding IT, 'smartness' and public benefit) but frequently not explained in much detail.

In practice the assertion reflects expectations that ANPR will -

• allow police forces (and by extension other law enforcement or national security agencies) to concentrate their resources and for example pull over specific vehicles rather than all vehicles or focus attention in a specific part of a road network
• automate some activity that would be overly expensive if undertaken on a manual basis and thus, for example, enable a reliable low cost traffic congestion pricing regime
• allow rapid integration of information that is useful to law enforcement personnel 'on the spot', for example that special caution should be taken in dealing with the driver of a vehicle that has breached a speed limit because that vehicle is flagged as stolen, was involved in a petrol theft or was recorded as being in the vicinity of bank robbery
• mechanistically identify vehicles and traffic patterns that would be 'missed' by human observers, described by one enthusiast as automatically and accurately finding needles with a particular characteristic in an ever-moving haystack
• offer 'remote' surveillance by law enforcement agencies of 'persons of interest' (such as suspected drug traffickers and terrorists) without the problems caused by use of implanted tracking devices and 'shadowing' by another vehicle.

Community attitudes to identification vary considerably. It is clear that individuals are often supportive of measures that are promoted as -

• inhibiting car theft (and increasing recovery of stolen vehicles)
• reducing the risk of carbombing and other terrorist incidents
• aiding action against drug trafficking, bank robbery, carjacking and kidnapping or
• merely reducing unnecessary disruptions, questions and searches by police (and thereby embodying traditional notions of privacy as "the right to be left alone").

However, some individuals and civil society organisations are wary of the potential to integrate ANPR data with other information collection and analysis activity to provide a panoptic sort - one that it is difficult to evade and indeed may not be apparent.

Others simply oppose ANPR on a visceral basis as a mechanism for speed fines and congestion management pricing.

     issues 

Use of ANPR systems poses a range of issues for officials and the community, including -

• fuzziness
• regulatory creep
• aggregation
• marketing
• privacy

     fuzziness

As the name suggests, ANPR identifies a plate - the registration number - rather than the vehicle as such.

It can be subverted by using a legitimate plate from another car, by using a fake plate or by obscuring a legitimate plate (eg with artfully applied 'mud', oil or even a plastic filter).

Responses depend on objectives in using ANPR. Subversion by muddying the plate and thus preventing capture of a usable image is arguably less serious in evading congestion charges than in evading apprehension for crimes of violence. An obscured plate might simply signal the need for verification by a patrol car (with traffic law typically featuring an offence of deliberately obscuring a plate) or recording by CCTV.

Using a plate that is not in the register and thus cannot be matched or that is a duplicate (with the same plate being captured from two vehicles in different locations at around the same time) would similarly signal a need for verification. Police forces that have adopted ANPR thus typically boast that the systems have helped to weed out fake and cloned plates.

Illicit use of a legitimate plate on another vehicle might not automatically generate a flag but would risky if a law enforcement agency is conducting spot checks or systematic checks of traffic at a particular point. The database reports that the plate belongs on vehicle type x; the observer can see that it is on type y and accordingly seeks clarification from the driver.

It is likely that ANPR will be supplemented by electronic vehicle identification (EVI) schemes, notably RFID tags built into the bodywork of vehicles in addition to tags on plates. Some schemes envisage that each vehicle will have a unique whole-of-life vehicle identification number (VIN) that is independent of the number plate and unaffected by changes in the ownership of that vehicle. Discrepancies betwen information provided by one or more tags would indicate a need for inspection by a human, inhibiting casual cloning of plates and illicitly swapping plates between vehicles.

ANPR identifies a plate rather than a driver.

The vehicle might for example be under the control of an individual as the owner of that vehicle and registrant of the vehicle. However, many vehicles are not driven by the registrant. A vehicle might for example be leased from a car/truck rental service, on a long or short term basis, or driven by an employee who is using a corporate fleet. A vehicle might be driven by someone who has been authorised by the registrant, for example a spouse, child, employee or friend. It might instead be driven by someone such as a thief who has no authorisation.

The relationship may not be determinable in real time although identifiable on a retrospective basis (eg when officials examine the records of a van rental company after a vehicle has been rented by a terrorist, kidnapper or other criminal).

Proponents have commented that 'you have nothing to fear if you have nothing to hide, asserting that 60% to 70% of unregistered and uninsured drivers "are actively engaged in crime".

     regulatory creep

An independent issue is regulatory creep, with ANPR systems being -

• initially deployed and accepted by the community for a specific reason (and regulated on that basis)
• subsequently discovered to have a range of uses that were not envisaged by the community at the time of implementation and thus are inadequately regulated.

Regulatory creep is an institutional imperative, with officials under pressure to make 'best use' of their resources and understandably eager to solve problems by using existing systems in new ways. It is problematical because their values may differ from those of lawmakers or the community, because news uses may be inadequately addressed through existing regulatory arrangement, and because there may be considerable uncertainty about what the nature of those new uses and any new regulation.

In the UK for example there have been conflicted statements about whether some ANPR systems are (or indeed could be) used to determine a vehicle's speed or its route. One spokesman thus commented that

These cameras are most definitely not enforcement cameras or speeding cameras. Their function is to provide police forces with intelligence for further action

Another spokesman, referring to intelligence, claimed that cameras would be used for public safety applications such as ascertaining the location of carjackers or suspected terrorists in real time. A colleague commented that "we only ever use the information for the investigation of crime" before noting that a UK force would provide information under a court order (although that apparently has not eventuated in divorce or other civil law disputes).

Debate in parts of the US has featured contention about how quickly congestion management systems would be used in surveillance of suspected drug dealers or merely the cheating spouses of the system operators.

     aggregation

Some critics recognise that government agencies have previously been able to gather information about traffic, including manual notation about the plates of vehicles that pause at a particular intersection or that are stopped by officials.

Those critics have however expressed concern because ANPR potentially allows such agencies to quickly build very large databases about -

• what vehicles are in use (ie on the road at a particular time, rather than at rest near the family home, a commercial parking facility or corporate park)
• where those vehicles have been and, by inference, where they are likely to be going
• how quickly they are travelling

Some concerns centre on provincial or national ANPR networks, with the UK National ANPR Data Centre for example facing criticism because it will centralise ANPR data from 43 police forces in England and Wales, drawing on several thousand cameras. In Australia it is likely that CrimTrac will provide a national clearing house that features information from discrete ANPR networks maintained by the Australian Federal Police, the state police forces and other agencies.

Criticisms reflect perceptions that large databases - particularly those accessed by a wide range of users, some of whom have inferior security protocols - are susceptible to abuse. One contact thus commented that

I might trust the Australian Federal Police but when I read year after year about corruption in some state forces and in vehicle registries I'm less inclined to trust them.

Other concerns centre on uncertainty about the accuracy of information held on ANPR databases. In the UK some of the more breathless media coverage has claimed

from the point at which a vehicle passes an ANPR camera, it takes four seconds for a police intercept team to receive data on whether a vehicle is stolen, has been involved in a crime, or is under surveillance. During this time, the number plate is checked for matches against the PNC database and several other intelligence databases including Revenue and Customs, the DVLA and Motor Insurance databases, allowing officers to identify vehicles that are not registered, taxed, insured or are without a valid MOT.

The quality of some of that data is poor, with recurrent criticisms for example that vehicle insurance databases are badly maintained and that promotion of ANPR as a mechanism for interdicting "uninsured and unregistered drivers" is thus problematical.

     marketing

Australia, along with the UK, does not regard vehicle registration details as publicly accessible information and unlike parts of the US thus does not publish its numberplate databases or sell them to data brokers.

Commercial datamining of vehicle registration databases is practiced in the US, for example to target offers from insurers and financiers, and in principle could be combined with information from private sector ANPR networks, with cameras capturing images of vehicles in public places.

A preceding page of this note highlighted the existence of private ANPR systems in Australia (for example to manage car parks and as the basis of 'drive off' blacklists in petrol stations). Given claims that those systems are outside the national Privacy Act 2000, because not featuring personally-identifiable information, they might be used for promotional offers when a vehicle recurrently visits stations within a networked chain.

     privacy

Much criticism of new technologies such as BPL relates to skepticism about whether real world performance matches hyperbole from boosters. In contrast, most critics of ANPR recognise that -

• the cameras can indeed collect data with a high degree of accuracy
• the data can be readily integrated with other information collections (including offender registers and watchlists) created by government agencies and private sector entities (eg credit card providers and vehicle rental services)
• subject to regulatory constraints, data can also be readily provided to third parties

They have accordingly asked whether ANPR erodes privacy and whether it subverts the civil society principle of 'innocent until proven guilty' by enabling pervasive surveillance of both passengers and drivers.

One critic sniffed that

traditional law enforcement is based on the assumption that the police only track and watch people where solid evidence of criminal activity already exists. It does not involve tracking everybody going about their normal business by using electronic tools such as CCTV, ANPR and systematic profiling that infers past deeds or future action on the basis of your residential address, work address, sex, how much tax you paid last year and whether you travel overseas.

Such criticisms centre on integration with other data sets, rather than mere collection of number plate information, and of action based on that integration.

Most official ANPR systems appear to be designed to accommodate access by a range of users and to enable exchange with different law enforcement/national security networks. The British ANPR system for example is reported as being compatible with the EU-wide Schengen Information System (SIS) that features criminal and 'wanted person' data from most EU states.

A response has been to restrict access to ANPR databases and use of ANPR data. In the UK for example information from the ANPR networks is regarded as personal data under the Data Protection Act that - formally - after 90 days can be accessed only for tackling serious crime.

   next page (ANPR in Australia)