When Bruce Schneier hears that millions of dollars are coming to Pittsburgh from the federal government to put surveillance cameras on our bridges, buildings and highways, "I just sigh and think, 'More money wasted,'" says the California security-tech expert. "What are the cameras supposed to do?"
The simple answer: They're supposed to protect our rivers and port facilities from terrorism ... even though those in charge of waterway security -- the Port of Pittsburgh Commission and the local Coast Guard branch -- can't cite a single threat in the past.
The cameras are also supposed to deter crime, says Mayor Luke Ravenstahl, especially when the city puts even more cameras into the neighborhoods. The Department of Homeland Security, through FEMA, is paying to put remote-controlled cameras on our bridges, license-plate-reading cameras on the highways, mega-pixel cameras atop the USX tower, and gunshot-detection cameras in a pair of city neighborhoods and Point State Park.
And that's just the beginning: The administration hopes to install yet more cameras in business districts, and in certain neighborhoods deemed high risk.
"What you want" from all these cameras, muses Schneier, "is for someone to say, 'Oh look, there's a camera; I'm going to quit my life of crime and get a real job.' Not, 'Oh look, there's a camera, I'll go around the corner.' Or for a suicide bomber to say: 'Oh look, there's a camera. Let's wave.'"
The federal government has provided $2.5 million to purchase the cameras, which comes from the Department of Homeland Security, and that money will be augmented by $862,000 in matching funds the city is pitching in. The money could have hired back a lot of the police officers Pittsburgh lost in previous years. But that's not where the funding is nowadays.
"I forget who coined the term 'the security-industrial complex,'" says Schneier, but he believes it certainly applies to the post-9/11 world, where companies are making huge profits from our fears, and politicians look good helping them do so.
If you believe there's another solution to genuine security issues, "you're a lonely voice," he concludes. "We live in a society where now, every security measure, no matter how stupid, we all accept. All we can do is keep complaining."
Or keep looking more deeply into the issue of whether cameras are really keeping us safer.
Pittsburgh Police Cmdr. Linda Barone, the department's point person on the impending surveillance system, did not answer multiple requests to speak about how the cameras might be integrated with future Pittsburgh police work. But cameras have, of course, been in use inside and outside businesses for years. And police Lt. Kevin Kraus is certain they have prevented crimes.
"They've proven very helpful in the past, very beneficial for capturing individuals," says Kraus, who has managed the Pittsburgh Police Crime Stoppers tip line for the past three years.
Often, in fact, the cameras work best when they are least effective at changing criminal behavior. "You still get people who come in with no attempt to disguise themselves" while robbing a store or bank, Kraus marvels.
But even these tried-and-true cameras can be unreliable. "Usually the images are usable," Kraus says, but "a lot of times there are surveillance cameras in use and they don't record.... Sometimes we find systems that are malfunctioning at the time of the actual crime."
It's even harder to know how well cameras work when they are installed across whole blocks or neighborhoods.
Press reports from across the country provide a murky picture of the cameras' usefulness. In January, for instance, the San Francisco Chronicle reported that 68 city-funded cameras perched above the city's "toughest street corners," costing $900,000, had aided a single arrest over two and a half years, in a city whose murder rate hasn't been so high in a dozen years. Four months earlier, the Chronicle had reported that 178 federally funded cameras aimed at the city's housing developments had failed to stem the high murder rate there. Yet the city is set to install 25 more cameras, costing roughly $250,000.
On the other hand, in 2007 alone, a surveillance program in Thunder Bay, Ontario, helped police defuse 78 "incidents" (presumably potential crimes), make 69 arrests and respond to 13 medical emergencies, at a cost of about $167,500 that year, according to a local account.
Here in Pittsburgh, the Ravenstahl administration points to London's 10,000-camera "Ring of Steel," and the 1,000-plus cameras now blanketing Chicago, as models for Pittsburgh to emulate. Indeed, Chicago police spokesman Cmdr. Jonathan Lewin says total crime is down 20 percent in Chicago since the cameras were introduced in 2003 -- including a 27 percent decrease in homicides -- and public demand for more cameras is high. "In 2007, there were more cameras installed than in the first four years of the program ... combined -- a pace we don't see decreasing," Lewin wrote in an e-mail response to City Paper questions. Chicago has counted 1,770 camera-aided arrests since the department began its tally two years ago.
"We believe technology, including the use of ... cameras, has played a role in this decrease," Lewin adds. Chicago police have fully integrated surveillance into their operations. The department can relay footage in real time to commanders, detectives and precinct houses -- even some squad cars.
But Lewin cautions that the cameras are "but one strategy in an overall effort to reduce violent crime," which included targeting gangs with more officers on the street and high-tech mapping of criminal activity. And, he adds, "no objective, academic analysis of the effectiveness of police surveillance cameras has yet been conducted in the U.S."
Indeed, an August study by the Northern California American Civil Liberties Union found that, of the 131 California cities surveyed about their surveillance practices, "no jurisdiction has conducted a comprehensive evaluation of the cameras' effectiveness." The ACLU cites three studies in the U.S. and Britain, in which convicted criminals ranked cameras low on the list of law-enforcement tools they feared.
"Other Western democracies are more interested in the research than [are] our own," observes Brandon C. Welsh.
At the behest of the Swedish National Council for Crime Prevention, Welsh, an associate professor of criminal justice at the University of Massachusetts-Lowell, analyzed 41 surveillance-camera studies from around the world. In October, he and a co-author concluded that "CCTV [closed-circuit television] caused a small (16 percent) but significant decrease in crime" on average in the locations studied, as compared to similar nearby areas not covered by cameras. "However, this overall result was largely driven by the effectiveness of CCTV schemes in [parking lots], which caused a 51 percent decrease in crime." And, he noted, all of the parking lots with crime decreases had also added fresh lighting and more security officers.
Welsh's analysis found an average crime decrease of only 7 percent in downtowns using cameras, a reduction he termed "small and nonsignificant" -- that is, conceivably due to chance.
Welsh's Swedish report was an updated version of an analysis he conducted for the British government in 2000. "The Blair government did not want to release our study," Welsh says of then-Prime Minister Tony Blair. "They had already spent hundreds of millions of pounds ... with very little to show with respect to crime reduction. And yet countries and governments continue to roll these programs out."
In fact, in the United States, the Homeland Security department is in the midst of a large national push to bring surveillance equipment seemingly everywhere, as the Boston Globe reported in August. That includes more populous cities like St. Louis and Madison, Wis., but also the 95 residents of Liberty, Kan., who are getting a camera for their park, and the 14,000 souls of Scottsbluff, Neb., who are now under constant watch as well.
In what appears to be a rare move, the Village of Fontana-on-Geneva Lake, Wis., actually rejected, on Feb. 4, a $12,000 Homeland Security grant, the local Janesville Gazette reported. (The idea of a surveillance system in this tiny hamlet "seemed to make a lot of people nervous," the village president told the Gazette.)
But few cities have the stomach to turn down free funds, especially with 71 percent of Americans wanting to see more surveillance cameras, according to an ABC News/Washington Post poll in July.
"Cameras capture the public's imagination. They feel like they should work," says Schneier, the California security expert.
Schneier says the way to reduce crime is often much simpler. "Lighting reduces crime!" he says, citing a 2002 British study that reported improving lighting could decrease crime by an average of 20 percent.
"Nothing beats policemen on the beat," Schneier adds, "and it's something our society has lost over the decades. That's real policing. It's old-fashioned, so it's less sexy. But [cameras] sound good" and they are "cheaper than a policeman."
Yet even in London's "Ring of Steel," a camera is no substitute for police work.
Last September, London's Evening Standard reported that "police are no more likely to catch offenders in [London] areas with hundreds of cameras than in those with hardly any. In fact, four out of five of the boroughs with the most cameras have a record of solving crime that is below average."
In the summer of 2005, Schneier points out, London cameras did "capture pictures of those London subway terrorists," whose bombs killed 52 people in July. Footage taken on June 28 of that year, just 10 days before the attacks, showed them practicing. But the images were discovered only after police knew who they were looking for, and only after a search through 80,000 hours of footage -- a job police didn't finish until three months after the attacks.
The cameras were "not much of a consolation to the victims' families," Schneier says.
The biggest problem with constant surveillance, say its critics, is the cameras' ability to turn ordinary life into episodes of America's Most Wanted.
A report last year, jointly issued by the Washington, D.C.-based Electronic Privacy Information Center and a similar U.K. group, labeled Great Britain and the U.S., along with Russia and China, as "endemic surveillance societies" for their increasing reliance on surveillance without a similar increase in privacy protections.
But at least Britain has laws governing the use and storage of video data, including an oversight office and privacy principles. The 94-page federal application for Pittsburgh's port-security grant, by contrast, doesn't even mention the word "privacy." Neither Pennsylvania nor the federal government, meanwhile, has rules to protect the public from feeling stalked.
"Most of us engage in a host of activities which are perfectly legal, for which we need to enter public spaces, but which we expect are private," notes Sharon Bradford Franklin. Franklin is senior counsel for the Washington, D.C., nonprofit The Constitution Project. She spoke to City Paper just after presenting the Project's suggested surveillance guidelines to a San Diego conference of a police chiefs' group.
Forty years ago, the U.S. Supreme Court ruled that there is no "reasonable expectation of privacy" in public, despite Fourth Amendment protections against "unreasonable searches."
"This is an area where the technology has really advanced a lot faster than the law," Franklin says. Those four-decade-old cases "were not confronting the kind of technology we have today or that is right around the corner. Law enforcement can create a digital dossier on individuals.... As a practical matter, that would violate most people's rational expectation of privacy."
The San Francisco-based Electronic Frontier Foundation has similar concerns. Says EFF senior staff attorney Lee Tien: "[H]ow would you like to have your own personal stalker? Someone who follows you around, records what you do every time you're on a sidewalk, or driving on the street, or in a shopping mall or other more-or-less 'public' space? They see you take your kids to school every day. They know your 'normal' routine. They know who you meet at Starbucks. They know when everyone's left your house. They see if you go to a ball game, a strip club, a bar, an anti-war rally, a gun show, an oncologist, a church, an AA meeting. This is 'connecting the dots' with a vengeance. Plus, of course it's not just the camera data" that law enforcement can and does collect. "They might also have your phone records, your Internet search records, your financial records, etc. Maybe you use E-ZPass; they know a lot about your driving."
Indeed, turnpike entrance and exit records have already figured in divorce cases.
Among the central issues for Pittsburgh's future system: Who will have access to the pictures, and how long will they be stored? The Constitution Project's model legislation calls for access only by trained law-enforcement officers, while acknowledging that criminal and civil court cases may see such data subpoenaed and revealed -- and that federal terror watch-lists may get access without even a court order.
The Project also recommends that camera images be stored no longer than seven days. But 30 days of storage "is what we'd like to see," says Henry Homrighaus, the San Antonio, Texas, security consultant helping to choose Pittsburgh's camera system.
"A lot of municipalities, airports, stadiums start out wanting to store 30" days, Homrighaus adds, "but end up reducing that to a couple of weeks" due to the volume of data.
A month's storage is "unnecessarily long," the Project's Franklin believes. But she was encouraged to receive a call from Pittsburgh's law department, asking about the Project's guidelines. She hopes the city uses "the principle of minimization -- you consciously minimize the impact on constitutional rights." For instance, after placing a camera on a bridge for port security, "that camera should not need to, and should not, pan-tilt-zoom to look at the apartment building next door."
The mayor's office did not answer CP's inquiries about possible privacy policies for the camera system. But city Councilor Bill Peduto has been working with the city's law and computer departments for the past six months to promulgate surveillance rules -- an effort he initiated in an amendment to council's bill accepting the Homeland Security grant. (Peduto himself has not been immune to the allure of the shutterbug: He has tried unsuccessfully to secure funding for traffic cameras aimed at those running red lights in dangerous intersections, a move he says was warranted by traffic deaths.)
"Before implementation of any system, we need to have a policy that clearly states who monitors the cameras" and where the cameras go, he says. While nothing has been finalized, he cautions that setting data-storage length is not a simple matter of minimizing. While privacy advocates may want to see camera hard drives wiped clean frequently, a longer record may aid a public claim of unjust arrest or other police misconduct, for instance.
(In fact, one area where surveillance cameras have proven their effectiveness is in police patrol cars. Reports the Seattle Post-Intelligencer: "The International Association of Chiefs of Police sponsored a study that found 93 percent of police-misconduct cases in which video is available result in the officer's exoneration.")
Thomas Nestel III, a former Philadelphia police inspector, examined surveillance practices in the top 50 U.S. cities for his 2006 master's thesis for the Naval Postgraduate School. Nestel, now the police chief in a Philly suburb, says his scariest finding was the lack of written surveillance policies. Only 6 of 17 cities using video monitoring at the time had any written policy governing the use and retention of video footage.
"I argued that any time you have a camera operation up and going ... there should always be an on-scene supervisor monitoring the use of cameras" to avoid abuse, Nestel says. Plus, it protects the police. "Anywhere where there's a possibility of a constitutional challenge, you have a written policy," he adds. "I think CCTV can be very valuable, but if you don't have safeguards, then the police are going to lose that tool forever. There's going to be a legislative response to that misuse. And it's going to be very restrictive."
Last summer, U.S. Sen. Joe Lieberman, of Connecticut, amended the Homeland Security appropriations bill to mandate that the department set national rules for the use of surveillance, including assessments of civil-rights concerns and alternatives to CCTV.
In the meantime, cities' rules vary widely, Nestel found. Routine storage of data, for instance, ranges from three days to "indefinitely." In Chicago, Mayor Luke Ravenstahl's proposed model, Cmdr. Jonathan Lewin writes: "Video that captures criminal activity is preserved for investigative purposes throughout the duration of the case. When not needed, video is erased within fifteen days."
While Nestel favors The Constitution Project's guidelines, Bill Peduto is so far a fan of rules put forth by the New York City chapter of the ACLU, which adds cautions against making an "unofficial archive" of tapes of specific lawful events, such as public protests, or making visual records of fliers or other materials protected by the First Amendment.
As Melissa Ngo, director of the Electronic Privacy Information Center's Identification and Surveillance Project, points out in a book chapter coming next month, "There is a right" to privacy in public -- "one of anonymity. We have the right to be unidentified in public, but this is not possible with constant camera surveillance.... [In public], unless a strange incident arises, we pass each other on the sidewalk and instantly forget the features of those who have just walked by. It is this inability to recall complete, specific data that leads people to expect anonymity in public places. No one expects to be constantly recognized and categorized while going about their normal lives, yet surveillance cameras record every detail of our interaction with the environment, making those details available for infinite replay and scrutiny."
But that, it seems, is where the future is headed.
This month, some of England's ubiquitous cameras have added both canned vocal warnings and speakers for police, watching live, to shout warnings at people. Somewhere far out of video range, George Orwell is thinking this is doubleplusungood.
Pittsburgh won't have shouting cameras -- yet. For now the priority is that the system "have the ability to be far-reaching" to see down the lengthy waterways and to "work in all lights," says Henry Homrighaus, the Texas consultant. "[I]f something crosses a line or isn't where it should be," the cameras will trigger a law-enforcement response.
But locally and across the country, technicians are rapidly increasing the power of surveillance technology. Soon, such systems won't just be able to watch us ... they'll be able to watch themselves watching us, and to figure out who we are.
Some of the biggest, and potentially creepiest, advances in surveillance technology are being made in the field of videoanalytics, which vision researcher Vladimir Brajovic defines as the science of "teaching computers to make conclusions about pictures."
Brajovic was formerly a student and professor at Carnegie Mellon University's Robotics Institute, trying to make the machines see better in natural light. Two years ago, he spun off his Bridgeville-based firm, Intrigue Technology, to develop surveillance cameras that can discern objects simultaneously in deep shadows and blazing light, conditions that leave current equipment blind.
The most dazzling sunlight, he explains, is a million times brighter than the darkest shade, but a normal camera can see only about a thousandth of that range. The human eye isn't perfect either, but the human brain can detect changes in light from one spot to another, making out shapes where cameras go blind. But with funding from the National Science Foundation and the Department of Defense, Intrigue is developing a camera that can discern patterns much like the human brain does.
Videoanalytics has existed for about 40 years, and the technology is still far from perfect. But, says Brajovic, "We've made good progress." Commercial software can already notice an object missing or suddenly present in a scene, or "if someone is fighting against the crowd. That guy may be suspicious. What is he doing?"
So far, the cameras "make too many mistakes," in answering that question, Brajovic concedes. But even so, places like "airports are huge" as a potential market.
And just over the Pennsylvania border, in Fairmont, W.Va., Mary Ann Harrison is developing the "Tactical Analysis of Video Imagery" (TAVI) project for use in Iraq. The system uses videoanalytics, "looking for immediate dangers" to help ward off insurgent attacks, explains Harrsion, the lead scientist in the West Virginia High Technology Consortium Foundation's Scientific Research Group.
TAVI will employ a wide-angle-lens camera to look for changes in a scene -- a package left behind, or people who are loitering or violating a perimeter. If it notices something suspicious, the first camera will ask a second camera -- a highly mobile pan-zoom-tilt model -- to get head-and-shoulder shots of any persons involved. TAVI will then compare those faces to a database of potential insurgents, Harrison says. Using face-recognition software, TAVI will either find a match to a suspect watch-list or "mark them as 'Unknown Insurgent," and zip the photos to soldiers in the field.
From there, TAVI will examine "anybody who was engaged in conversation with this person for a long time or walked with this person for a long time," plus study what materials they purchase, toward what possibly nefarious end.
"On the basis of what connections and of what type" are observed over the long term, Harrison says, "you can infer leadership," as well as the importance of the entire group. "Is this cell a really minor cell -- but do they have an attachment to a more major target?"
Harrison's work, like that of Brajovic, is funded by the Department of Defense. But as with so many DoD projects, this one will inevitably migrate to civilian use.
"It doesn't take much imagination" to see it applied to law enforcement, Harrison says. "There are a lot of civilian applications," from tracking gangs to spotting "suspicious behaviors" in an airport security line.
Pittsburgh's system, once established, will resemble the TAVI system in some ways. It will use secure Web pages to display, and hard drives to store, camera footage in digital form, and it will employ some form of videoanalytics -- though Homrighaus says "at this point," it won't feature face-recognition software.
Meanwhile, police and private citizens have begun using YouTube as a way to distribute surveillance images, making us all potential police and potential suspects. That's not even including ubiquitous cell-phone cameras, or the street views of Google Earth.
Ten years ago, New York City's Surveillance Camera Players, a performance group, tried to map all of Manhattan's surveillance monitors, public and private. Today, says Bill Brown, a group co-founder, part of the organization's task is trying to keep track of the Homeland Security dollars that are expanding the city's surveillance apparatus even further. The city is supposed to get thousands of new cameras, both in the subways and to build the city's own "Ring of Steel."
"Here in New York City, it's fast and furious," Brown says. Since 9/11, there had always been talk of putting in such measures. "Now it's finally taking place." Terrorism fears are being exploited, he says, which is "just turning safety into a business."
"There's been no time for people to discuss the risk to put in these cameras," Brown concludes. "And that's precisely the reason we should slow down and talk about it."
"With civilian applications there will be privacy concerns," allows Harrison, of the Scientific Research Group. "Already in our daily lives we're being monitored constantly. There's just cameras everywhere. Video is probably the only sensor that can monitor human behavior without people being aware of it and giving permission."
