Thursday
Mar172011

"Drones Set to Invade National, State Parks" on Wired's Danger Room

"When I was a kid going to summer camp in New York’s Adirondack Mountains, I counted myself lucky if I saw a black bear once or twice in a season. But campers may soon be able to regularly see something bigger and badder when climbing the High Peaks: Reaper drones flown by the New York Air National Guard’s 174th Fighter Wing based in Syracuse, New York..."

Read the full post at Danger Room.

Thursday
Nov012007

“The Effectiveness of CCTV Surveillance as an Anti-Terror Intelligence Tool: The 7/7 London Train Bombing Attacks, Today, and Tomorrow”

Background

This paper was written in the fall of 2007 in Georgetown University’s School of Foreign Service Security Studies Program. The assignment was to write a short analytical brief (no more than five pages) on a current topic in the area of technology and intelligence.

Abstract

The proliferation of CCTV surveillance technology has grown throughout the past decade, especially in the wake of Al Qaeda directed or inspired attacks in New York, Madrid and London. Proponents of this trend claim that these networks, combined with the new data mining and management software, create effective tools for both anti-terror and anti-crime operations. Detractors point to the failure of cameras to thwart major attacks in an anti-terror role as well as their limited value in an anti-crime role. This paper will assess the effectiveness of CCTV surveillance as an anti-terror intelligence tool in the 7/7 London Train Bombing attacks, and examine CCTV surveillance today and in the future.

The 7/7 London Train Bombings

On 7 July 2005 three bombs were detonated on the London Underground and a fourth was detonated on a London bus. The attacks killed 52 and injured over 700. “This is the worst terrorist attack in Britain,” reported The Times, “The previous highest toll was in the 1974 Birmingham pub bombings, which killed 21.”1 A series of follow-on attacks planned by a different but related group was subsequently thwarted.2 Although direct links to Al-Qaeda were not clear, both the successful 7/7 bombings and the failed follow-on attacks bore the hallmarks of Al Qaeda involvement that had been demonstrated on 9/11 in the United States and in the Madrid Train Bombings: massive simultaneous attacks against vulnerable civilian infrastructure.

CCTV surveillance prior to the bombings: anti-crime versus anti-terror approach

By 2005, the use of CCTV surveillance had reached a high level of coverage in the United Kingdom since its introduction in the late 1990s. However, even in the aftermath of the 9/11 attacks in the United States, the focus of most CCTV surveillance was as an anti-crime tool as opposed to anti-terror tool.3 This is an important distinction for two reasons. First, although both criminals and terrorists commit criminal acts, their motivations can be different, which can cause different behavior patterns to produce similar results. Second, the response of the state to crime and terrorism varies and has changed over time.

For example, in the United States pre-9/11 state response to terrorism was framed in terms of law enforcement: terrorists were criminals in systems of states. They were often arrested, tried and convicted using the same systems the state employed against criminals. In this framework, anti-crime and anti-terror policies could and did share a number of common features. Post-9/11, however, the United States response has been framed in terms of war-making: terrorists are enemy combatants in a global, sometimes seemingly borderless battlespace. They are targeted for destruction and sometimes capture using the same systems the state employs against armed enemies (historically national armies). In this framework, anti-crime and anti-terror policies cannot and do not share as many common features.

A critical difference in the approach to information gathering in an anti-crime posture and an anti-terror posture is the end result of the intelligence it creates. In an anti-crime system, information is collected primarily to gather evidence to catch and prosecute criminals after a crime has been committed. Even though there are some moves afoot in the United Kingdom today to use CCTV surveillance in a more pro-active crime prevention role, cameras have generally been used in this fashion.

In an anti-terror system, by contrast, a major goal of intelligence collection is to prevent attacks before they happen. This means that the system must detect the indications and warnings of an attack, relay the information to active elements of the organization that are authorized and enabled to act, and then kill or capture or disrupt the terrorists before they strike.  This distinction is especially important when looking at suicide bombings. If the attacker does not expect to survive, then he or she will most likely not be deterred by a system that is designed to capture and imprison him or her after their crime has been committed.

Effectiveness of CCTV surveillance as an anti-terror intelligence tool before the bombings

Clearly, CCTV surveillance was insufficient to prevent the 7/7 bombings. In this respect, its value as an anti-terror intelligence tool was low. Although 9/11 and the Madrid Bombings had clearly show that an attack like the 7/7 bombing’s was possible, British officials were caught by surprise.4 Certainly, CCTV may have had an impact on anti-terror operations that remains classified. But for the most part, its direct impact appears negligible, at least on the surface.

It can be argued that using CCTV surveillance in an anti-crime role still has indirect effects on countering terrorist operations, since terrorists sometimes also conduct criminal acts. However, the effectiveness of CCTV as an anti-crime tool has been questioned by numerous studies both in the United Kingdom and the United States.5

Effectiveness of CCTV surveillance as an anti-terror intelligence tool after the bombings

In-so-far as CCTV surveillance was able to help security forces conduct a forensic analysis of what happened, there was some value in that analysts could build a profile to use in screening for future attacks by other cells. Such work was essential in piecing together information that lead security forces to the second cell later in July 2005. In this regard, CCTV surveillance can play a role in anti-terror intelligence gathering. In general, after the 7/7 bombings, the volume of CCTV surveillance in the United Kingdom has increased.6 But based on its inherent limitations, it is unclear that an increased volume of coverage will result in an increased amount of actionable intelligence.

The future of CCTV surveillance

However, as seen in other technical intelligence collection disciplines (COMINT, IMINT, etc.) a greater volume of data does not necessarily lead to a greater amount of information, and a greater amount of information does not necessarily lead to a greater amount of intelligence. This is the classic “signal to noise” problem: how does the system of people and machines distinguish that which is important from the vast amount which is not.

This problem is even more acute when one attempts to use information collected from CCTV surveillance to create indications and warnings intelligence of events before they happen. A root problem of all technical collection systems is that as the number of sensors and the associated amount of data increases, the number of operators to analyze data does not necessarily keep pace.7 And with sufficient numbers of operators (and the necessary training and infrastructure to support them), it is still difficult to create predictive intelligence. Even today, areas that are heavily monitored by human operators are still hit by terrorist attacks.8 The essential problem is that a video image of a person walking down the street doesn’t tell even a well-trained operator where a person is going or what they are going to do when they get there. In cases where the security forces do conduct aggressive profiling of potential terrorist targets, there are still a large number of “false positives” (or errors of commission) for every successful prediction.9

The dream of technologists is to increase the value of CCTV surveillance systems for both anti-crime and anti-terror intelligence operations by creating systems that can detect behavior.10 This will entail the development of pattern recognition systems capable of scraping raw CCTV data feeds and creating information that can be fed to data mining systems and used as queries to databases. The complexity of performing these operations simultaneously on real-time data from literally thousands of feeds will be enormous. Despite the enthusiasm of the technical community, this day is a long way off. Although parts of the total system have developed in the past decade, we are still far off from the integrated system that can automatically tie enough information together to detect a terrorist in time to prevent an attack.

Conclusions

•      CCTV surveillance did not significantly contribute to preventing the 7/7 London Train Bombings, although it did help prevent follow-on attacks.

•    CCTV surveillance is better suited to anti-crime than anti-terror intelligence gathering (although its value as an anti-crime tool may also be low).

•    The effectiveness of CCTV surveillance in either an anti-crime or an anti-terror role is decreased without adequate investment in infrastructure and human capital.

•    There is a significant amount of technology development necessary to realize the dreams of creating CCTV surveillance based behaviors prediction tools.

 


1 “Britain ‘defiant’ as bombers kill 52 in attack on the heart of London.” The Times (London). 8 July 2005: Home News 2.

2 Beaumont, Peter et al. “Police swoop on capital’s estates as hunt intensifies: Two held as armed officers raid flats; Detectives seek link with African cell; 72 hours that shook London.” The Observer. 24 July 2005. News, 2.

3 Or simply as an anti-congestion tool: “There are more than 6,000 CCTV cameras monitoring every platform on the system, although Mr. McManus [head of policing and security for London’s Underground] admitted that staff were more concerned with preventing dangerous overcrowding than spotting potential suicide bombers.” (Tendler, Stewart, Webster, Ben and Evans, Michael. “Security officials ask: Could it happen in Britain?” The Time (London). 12 March 2004. Home news, 7.) An earlier article notes that “Mike Brown, London Underground’s customer services director said that a spate of newspaper and television reports had blown the danger [of possible terrorist attacks on the Tube] out of all proportion” and that “…any measures taken to protect the network would need to be balanced against the need to move three million passengers each day as efficiently as possible.” (Webster, Ben and Henery, Michelle. “Fear of terrorism attack on Tube ‘is unwarranted’.” The Times (London). 19 November 2002. Home news, 6.)

4 “The Home Secretary, Charles Clarke, admitted that the London bombings had come 'out of the blue'. The security

services, indeed, had recently lowered their estimate of the threat from 'severe general' " the third highest level " to ‘substantial'. In the wake of the attack it was immediately lifted to 'severe general' again, but it was clear that there had been no warning.” (“ATTACK ON LONDON: THREE DAYS AFTER THE SECURITY SERVICES 'LET IN A GOAL', THEY ARE STILL IN THE DARK.” The Independent on Sunday (London). 10 July 2005. News, 6-7.)

5 Gill, Martin and Spriggs, Angela. Assessing the impact of CCTV; Home Office Research Study 292. Home Office Research, Development and Statistics Directorate. 2005. Welsh, Brandon C. and Farrington, David P. “Surveillance for crime prevention in public space: results and policy choices in Britain and America.” Criminology & Public Policy. Volume 3, 3, 2004. Both studies find that lighting may have as much or more effect on crime than CCTV surveillance.

6 Today there are a reported 4.2 million security cameras in the United Kingdom today, or one camera for every 14 people in the country. It added “…recent studies found that every British citizen is caught by a camera's lens 300 times a day.” (“Big Brother is Watching.” British Heritage September 2007, Vol. 28 Issue 4: p.6.) Surveillance camera use is also increasing in the United States. The American Civil Liberty Union of Northern California notes that in California: “37 cities have some type of CCTV surveillance program. 18 cities have significant CCTV surveillance programs of public streets and plazas; an additional 10 jurisdictions are actively considering such expansive programs. 18 cities have systems in which the police actively monitor the cameras.” (Schlosberg, Mark and Ozer, Nicole A. Under the Watchful Eye; The Proliferation of CCTV surveillance Systems in California. American Civil Liberty Union of Northern California: August 2007: p.2.) The New York Civil Liberties Union reported that there were “nearly 4,200 surveillance cameras… below 14th Street [in Manhattan], more than five times the number counted in 1998.” (Barron, James. “Civil Liberties Group Worries as City’s Electronic Eyes Multiply.” The New York Times. 14 December 2006: B03.) And in August 2006 47 cameras were being installed throughout Washington D.C, while about 80 cameras have been set up in Baltimore. (Cella, Matthew. “Spy cameras fail to focus on street crime.” The Washington Times. 13 August 2006: A01.)

7 For example: it is estimated that the average Briton is seem by camera 300 times a day. (British Heritage.) But how does that translate into tasks for operators? With a 2007 mid-year population of 61 million (“Data Finder.” Population Reference Bureau. 27 September 2007. <http://www.prb.org/DataFind/datafinder7.htm>), 300 views a day creates over 1.8 billion views that humans need to process, or nearly 20,000 per second. With a combined police force of around 140,000 (“Police number reach record high.” Criminal Justice System. 26 July 2005. 27 September 2007. <http://www.cjsonline.gov.uk/the_cjs/whats_new/news-3184.html>) this means that if every police employee in the United Kingdom stayed up 24 hours a day they would have to review one image every seven seconds (on more manageable eight hour shifts, this becomes a slightly more hectic two and a third seconds per image). Clearly, these numbers are inhuman, and police forces that use CCTV surveillance readily admit that they come nowhere close to reviewing everything the camera sees: “’We don't have time to sit and monitor these cameras at all times,’ [Police Capt. Bob Keyes of the Clovis, California police department] said. ‘There should be no expectation from the public that people are watching.’” (Bulwa, Demian. “Future fuzzy for government use of public surveillance cameras; still, some Bay Area cities hope to follow Clovis' lead.” The San Francisco Chronicle. 23 July 2006.)

8 “Last month, a Palestinian with bombs hidden under his clothing boarded a bus in Israel as a police surveillance helicopter hovered directly overhead. As the officers watched from above, he blew himself up and killed 19 people.” (Goodheart, Andy. “Public cameras accost privacy.” USA Today. 22 July 2002. 26 September 2007. <http://www.usatoday.com/news/opinion/2002-07-22-ncguest11_x.htm>)

9 Fulton, Wayne. Personal Interview. 24 September 2007. Mr. Fulton, a private security consultant with extensive living and working experience in Israel and South Africa, spoke to me about undercover surveillance teams in Israel.

10 “CCTV surveillance recordings from the Underground have been key in identifying the perpetrators. Yet given the possibilities of artificial intelligence, it’s compelling to consider how a video camera infrastructure could be put to more advanced work in identifying and preventing terrorist threats.” (David, Mark. “Electronics can protect subways against terrorism.” Electronic Design. 04 August 2005: p.19.) Specifically, “The current shift from analogue to digital, computerized CCTV systems also supports expansion. Digital, algorithmic techniques, like those currently being used in the City of London’s ‘Ring of Steel’ system and the system in Newham, East London, allow much larger systems to be automatically monitored because the systems can be programmed to automatically search for ‘abnormal’ or ‘unexpected’ events, behaviours or even people. In the City of London, cars moving the ‘wrong’ way down a street automatically trigger cameras to monitor the scene. Additionally, by linking digit al CCTV with image database technology, ‘algorithmic’ CCTV systems can be programmed to automatically scan for specified faces or car number plates or to ensure that people are where they ‘belong’.” (Graham, Stephen. “CCTV: the stealthy emergence of a fifth utility.” Planning Theory & Practice. August 2002, Vol. 3 Issue 2.)

Page 1 2 3