Category: videosurveillance

Blackberry and the London Riots

I’ve been in the papers and on radio and TV a bit in the last few days here in Canada, talking about the London Riots, both as a ‘token Brit’ and a surveillance expert. I’m happy to talk about my feelings as someone from Britain and I’ve made it clear to people that I am neither a technical nor a legal expert, but the conversation inevitably ends up in those domains and others which are really outside my expertise – and I’ve had to be careful what I say.

I’ve generally stuck to three lines:

1. That these riots don’t provide simple moral lessons, they are neither politically-motivated or just about ‘crime’, but they do have roots and implications which are profoundly political – this is about consumerism, class, inequality and exclusion.

2. That you can’t blame Blackberry. That’s like blaming the postal service for hate-mail. The problems for RIM here are twofold: bad public relations from being associated with rioting, and how much it is prepared to sacrifice the privacy of its users to help UK police in an effort to counter the bad PR.

3. That all the UK investment in video surveillance didn’t help stop these riots (see my previous posts).

People like Chris Parsons are the kinds of people that the media need to talk to about the technical issues, and there’s a really fantastic and detailed post from his blog here on Blackberry and security and privacy issues. On legal issues, there’s no-one better than Michael Geist on things like lawful access. His website is here. Michael writes a regular column for the Toronto Star and I was quite amused that when the Star called me yesterday, I had to remind them to talk to him about lawful access issues! The best sociological piece I have seen on the causes is from Zygmunt Bauman.

That said, here’s some links – There’s a podcast here on the Financial Post, which also has a good discussion with Tamir Israel of CPIC.

On the more social side here, syndicated in lots of local and regional papers.

And the usually strangely edited piece in my local paper, the Kingston Whig-Standard, here, also featuring my colleague, Vince Sacco.

London Riots and Video Surveillance, Pt.2

My last post was about the lack of any apparent deterrence of rioting from CCTV. However that’s not to say that video surveillance is proving of no use to the authorities. However the way it is being used says a lot about both the limits of CCTV and the general problem of analysis of video images.

As part of ‘Operation Withern’, the investigation into the rioting, the Metropolitan Police have set up a special section of their website, London Disorder Images, as well as on Flickr, which is essentially crowdsourcing the identification of suspects. Despite being the most well-resourced police force in the UK, the Met lacks the resources, time and expertise to analyse and identify everyone it wishes to identify itself, and with widespread popular anger about the riots, they are banking on opening up the process of surveillance and identification as being more efficient and effective – and they may well be right.

Of course, with the problems of lighting, angle, distances, and image quality, the images vary in identifiability – and bear in mind that the few posted so far are probably amongst the best ones – and no doubt there will be many misidentifications. And, in addition, hundreds of people are already being processed through magistrates courts without much need to video evidence. But it is a tactic we are seeing more and more in many places (e.g. Toronto, following the G20 disturbances).

London Riots and Video Surveillance, pt.1

 A really interesting map on the website of the US monthly, The Atlantic, illustrating the relationship between density of video surveillance cameras (CCTV) and recent incidence of rioting in London. There are many things one can get even from a simple map like this. It’s worth noting in particular that Wandsworth and Harringey are the residential boroughs with the highest concentration of CCTV, and have been hit by rioting. There are also places with both greater and less than average density of CCTV which have not had rioting.
 
Whilst you have to be careful not to mistake correlation for causality, and bearing in mind that this is not a statistically tested verdict, the main tentative conclusion one can draw is that there seems to be no relationship between the presence and density of CCTV and the occurence of rioting. This might seem like  a fairly weak statement, but it is yet more evidence that CCTV has little deterrent effect on crime of this sort (and of course, the rioting is not only explicable as ‘crime’ anyway).
CCTV_boroughs.jpg
 

Security systems and trust

Sometimes, little local stories give us the best insight into what living in a surveillance society is really like. This one is from a school in Virginia, USA. According to the local newspaper (via BoingBoing) a middle school student was suspended from school for opening the main door for a women who they knew who was unable to press the entry button because they had their hands full. The reason given by the school auhtorities is that the school has a secure entry system, in which people are supposed to press the entry button, look into a camera, and request entry. The student was suspended on the grounds that they were all supposed to know the rules, and that these rules were potentially of vital importance.

However this security-bureaucratic reasoning misses the key point that the child knew the adult concerned. Whilst security and surveillance systems are at least in part designed to respond to a supposed decline in social trust and an inceased ‘threat’ (which is very poorly supported by evidence anyway), there is good reason to suppose that placing what were previously matters of social negotiation into the hands of such ‘systems’, ‘rules’ and ‘technology’ further damages social trust.

Many questions then arise: what is this school, through this action and these systems, teaching kids about society? That security comes above all else? That no-one can be trusted? And that individual decision-making or social interaction is better replaced by impersonal systems? Surely, if education is the basis of the future of society, then what should be taught are the opposite lessons. This kind of subordination to systems is a form of training, of disciplinary control, not learning and education.

 

UK consultation on CCTV: a weak brew?

The UK government has released a consultation document on a ‘Code of practice relating to surveillance cameras’ (CCTV). The closing date for comments in May 25th.

I will go through the document in more detail but there are several initial things to note here:

1. I am interested first of all in the fact that the camera systems are refered to as ‘surveillance cameras’ rather than ‘security cameras’ or ‘safety cameras’ as in many situations I have encountered around the world.

2. This is merely a step toward a state code of practice. The government had promised to ‘regulate’ CCTV, and what many people might have legitimately expected from such a promise was legislation, in other word a statutory footing for surveillance cameras and legal controls. A code of practice is very much at the weak and volunteeristic end of ‘regulation’ if it is regulation at all. The proposed Code itself is really quite weak and presaged on “gradually raising standards to a common level.” with nothing that is mandatory.

3. The document proposes another ‘Commissioner’ to govern surveillance cameras, a ‘Surveillance Camera Commissioner’. This government, despite its avowed attempt to reverse the proliferation of Quangos, seems to want to create another one. One would think that this would naturally fall under the remit of the Information Commissioner, but it appears that the Tory attacks on the ICO (which have been going on in newspapers like The Times for some years and have now spread to other libertarian groups) have been having some effect. Does Britain need another Commissioner in the area of information, surveillance and privacy? I don’t think so. I think we need to clarify the roles of existing Commissioners, and reduce their number – provide adequate budgets and better guidance and division of labour. I suggested a few weeks ago that splitting the ICO into a Surveillance and Privacy Commissioner (which would incorporate the data protection function and absorb all the existing micro-commissions like Surveillance, Interception of Telecommunications and now this new proposed Surveillance Camera Commissioner) and a separate Freedom of Information Commissioner, would be the best solution.

4. The consultation document acknowledges that camera surveillance has increased too rapidly in Britain and has eroded privacy and been overly intrusive. That’s a start. However it also hedges this quite strongly by saying that the government does not intend to limit law enforcement’s abilities. I am not sure the two things are compatible – but I will have to examine the proposals in more detail.

5. The document acknowledges that “CCTV does not always provide the benefits expected of it” but explains this as largely down to technical and operation reasons rather than anything more fundamentally problematic. This is not necessarily justified by evidence or particularly insightful.

6. The document acknowledges that Automatic Number Plate (Licence Plate) Recogntion (ANPR / ALPR) is largely unregulated too and that it connects to all kinds of databases, yet proposes little more than auditable data trails.

7. The document mentions both flying drone cameras / Unmanned Aerial Vehicles (UAVs) and helmet-mounted cameras, but assumes mistakenly that these are ‘niche and novel’. If this can still be said to be true, it will not be for much longer, and the document is overly dismissive of the immediacy of this issue.

8. The document is way too cautious and has the fingerprints of a ‘Sir Humphrey’ bureaucratic avoidance of anything that might ‘frighten the horses’, motivated as it claims to be by “the wish to avoid imposing unreasonable or impracticable bureaucratic or financial burdens on organisations” and recommending “an incremental approach.” It is too late for incrementalism, about 20 years too late in fact.

At first glance, the consultation document appears to be a rather weak brew rather than the strong medicine that is required.

The Expansion of Video Surveillance in India

A recent market analysis (which contained many predictions, more of which tomorrow) identified India as one of the world’s fastest expanding video surveillance or Closed-Circuit Television (CCTV) markets, and the coverage of policing plans in the Indian media over the past couple of years would seem to confirm this. In particular, in the wake of the terrorist attack on Mumbai, authorities in all major cities have been pushing ahead with the intensification of security and surveillance measures. This is part of a more general expansion of surveillance in all areas of Indian governance, some of which, like the new biometric census and high-tech border surveillance and UAVs, I’ve mentioned here before.

Cities such as Chennai have announced plan for 10,000 cameras across a range of settings (interestingly in this case, ‘marriage halls’ were one of the first locations to get CCTV – perhaps someone can enlighten me as to why this would be – along with state banks and major malls) and the police chief is quoted as saying he wants “the whole city covered by CCTV.” Delhi is combining a massive expansion of CCTV with increasing numbers of police officers on the streets, so this is not a case of an inhuman technological gaze replacing the neighbourhood police officer. And here, as in the state of Gujarat, in cities like Ahmedabad, the road network is a particular priority with Automatic License (or Number) Plate Recognition (ALPR/ANPR) systems and cameras being installed on all major roads. This ‘Intelligent Traffic Management System’ (ITMS) is designed to be multipurpose and address security, traffic and emergency requirements.

The diffusion of CCTV to more remote and peripheral areas has also been remarkably quick. Just recently, the northern Haryana region has also announced a huge CCTV installation of around 5000 cameras in eight cities, which will be targeted at “shopping malls, main market, major traffic points and escape routes in these cities” – an interesting turn of phrase, which almost seems to portray the city as a prison. Just as in the major urban centres of the country, here too the new systems will employ analytics including movement recognition.

This expansion has not gone unchallenged – see this debate over some of the Chennai systems – but the debates seem rather lifeless and complaints seem to be limited to hoping that there will not be ‘abuse’ of the camera systems by police, and commenting on the lack of any regulatory body for video surveillance. Nor has it all been smooth in technological terms. The Delhi expansion of CCTV builds, as in many cases, from the security upgrades for a ‘mega-event’, in this case the Commonwealth Games in 2010. However, as with much of the infrastructure for these games, there were reports of systemic failure, if not a total lack of functionality from day one. The cameras for the event were apparently poorly calibrated and made watchers dizzy an in some cases, installed where no view could be obtained. It is also not the case that what many nation’s security authorities would consider to be priorities for video surveillance have actually already been covered, even where there has been a demonstrable threat: for example, it is only now that CCTV is being installed at Mumbai’s Chhatrapati Shivaji International Airport, which apparently had no CCTV at all prior to this.

Overall, there appears to be strong media backing for a combined state and private sector discourse that emphasises CCTV not so much as a protection against terrorism (though that is clearly present) but as an unquestionably ‘necessary’ or even simply ‘natural’ component of progress, economic development and modernisation. Consider, for example, this description of the new “shining steel” Metro system in the high-tech and global information economy service-centre region of Bangalore, where “automatic fare collecting gates, metal detectors, CCTV cameras and voice announcement systems” were all of a piece along with the announcement of the new ‘signature tune’ for the public transit network. And see also this rather peculiarly de-politicised description of the history (and future) of policing technology in India, written by a former senior officer from Kerala state, in which the British colonial imposition of fingerprinting in India is portrayed as a collaborative advance and in which, of course, CCTV is pictured as part of a similar and apparently totally necessary new series of technological advances designed to drag Indian policing out of a ‘medieval’ period.  At the same time, however the historic (and largely colonial) legacy of a slow-moving, fragmented and conflict-ridden bureaucracy is still resulting in a very uneven diffusion of video surveillance across this enormous country.

What’s Wrong with Video Surveillance?

Occasionally, you need to simplify and clarify. Someone asked me the other day, “so, what’s wrong with CCTV anyway?” Here was my quick answer.*

1. Does CCTV prevent crime?

The prevention of crime was the main rationale for CCTV in Britain back in the early days in the 1990s, and this rationale is still the main one currently in the USA…

But meta-evaluation of valid studies of CCTV by Welsh and Farrington, recently published as Making Places Safer by OUP, shows the following: that studies can only show a positive correlation between reduction in rates of crime and the installation and operation of CCTV in limited situations, namely in car parks and the like. This is because car theft is a more rational form of crime (the perpetrators are often professional criminals and they do not want to get caught). Most crime, especially street crime and violence is not so rational. People do not generally look up in their violent drunken haze and think ‘ooh, there’s a camera, better not kick this guy’s head in’.

According to Martin Gill and co.’s work evaluating 14 schemes across the UK, only 1 resulted in a clear reduction in crime over the longer term. CCTV can have temporary effects in reducing crime (and police studies always seem to be done in these early months and hence are very misleading), but over the years after installation, unless other things are done, the crime will return to similar levels. It’s those other things that are done – more community volunteers, neighbourhood watch schemes, better street lighting, economic regeneration – that make the difference to crime rates. People who think they ‘know’ it’s down to CCTV are just looking at A and B and thinking changes to A must be a result of B, without considering C, D and E…

What can be useful in this regard, knowing that temporary reductions can be made, is to use CCTV in targeted, temporary and flexible manner – i.e. if you are going to have video surveillance at all, make it moveable and used to target specific areas where there have been sudden increases in crime.

So, so much for prevention…

2. What about solving crime? Surely CCTV gives us lots of evidence?

Well, not as much as you might think. The biggest study of street robbery and CCTV in London (the city with the highest density of cameras in Europe), commissioned by the Metropolitan Police, showed that only 3% of such crimes were solved using CCTV . And, figures released in 2007 through Freedom of Information Act requests, showed that 80% of crime in London still goes unsolved even with this infrastructure.

3. But at least someone’s looking out for us – right?

Studies of control rooms show that the professionalism and seriousness of operators is increasing but there is still evidence that more time is spent on anti-social behaviour and dealing with ‘unwanted’ people than the potential for serious crime, particularly in shopping districts. Who is watched and why is also complicated by family and social connections, especially in smaller towns. CCTV systems are also increasingly difficult to watch as the numbers of cameras and screens increase; there aren’t enough staff, however well-trained they are, to do a really efficient job in most cases and computer analytics are not good enough (yet).

In addition, there is the growing issue of cost. There were originally subsidies for installing CCTV in Britain in the 1990s, but running costs, maintenance and replacement have to be covered by the operators (usually Local Authorities) and there is an ongoing row going on behind the scenes between LAs and police in the UK about who pays for it.

Now costs are starting to bite, exacerbated by recession and new Tory efficiency savings, some local authorities have even started to either combine their monitoring with others – meaning even more distant and less efficient watching and in some cases have stopped watching the cameras live at all (in many countries this dead recording is normal anyway).

4. What about the Courts?

The only undoubtedly positive effect seems to be that it encourages criminals who are caught to confess and plead guilty, which saves court costs and time – although of course, guilty pleas mean that the criminals are punished less and out of prison quicker (if they go in at all), which might be felt not to be an advantage by some!

Summary

Video Surveillance, particularly fixed CCTV,  is expensive, inefficient and has all kinds of negative social side-effects. Public money would be better spent on improved street lighting, schemes for community involvement and economic regeneration. CCTV certainly isn’t a ‘Panopticon’ because actually it doesn’t actually ‘see’ very well at all nor does it actually seem to alter behaviour as much as states would like in itself, but it does appear to contribute to the decline of social trust and decreasing personal responsibility, which are partly at least to blame for the problems CCTV is supposed to solve, and all of which would be more likely to increase with other solutions.

*In most ways, this answer is not really ‘mine’ – it’s the distillation of many other people’s work, some of whom are mentioned here, some of whom like Clive Norris, Mike McCahill, Will Webster, Pete Fussey and Gavin Smith, are not… anyway, they know who they are! Thank-you all…

The Total Surveillance Society?

Advanced visual surveillance has become prevalent in most developed nations but, being restricted by inconvenient things like democracy and accountability (even if they are not as strong as some would like) and police and local authority funding, such surveillance remains patchy even where it is widespread.

The Chinese state, however, suffers from none of these inconvenient restrictions. Free from democracy, accountability, and with a buoyant economy still largely connected to the Communist Party, it is able to put in place surveillance systems beyond the wildest dreams of the most paranoid western administrators. The target of the new wave of surveillance is internal political unrest, particularly in separatist Tibetan Buddhist and Muslim areas of the massive nation.

Associated Press is reporting official internal announcements about how Urumqi, capital of the Uighur Muslim area of Xinjiang, which saw extensive anti-government protests last year, will be blanketed by surveillance systems. According to the report:

  • 40,000 high-definition surveillance cameras with riot-proof protective shells have already been installed in the region, with 17,000 in Urumqi itself
  • 3,400 buses, 4,400 streets, 270 schools and 100 shopping malls are already covered
  • the aim is for surveillance to be “seamless”, with no blind spots in sensitive areas of the city (and this includes in particular, religious sites)
  • 5,000 new police officers have been recruited

This is part of a wider ‘Safe City’ strategy – in this context, even more of a euphemistic description that the same words would be in the west – that will see 10 million cameras being installed across the country. Ths numbers keep growing all the time: the last time that I reported on this, the estimate was less than 3 million ! IMS Consultants last year estimated that the Chinese video surveillance market was $1.4 billion in 2009, and that this will grow to over $3.5 billion by 2014. China is now the single largest market for video surveillance in the world.