The good and bad of the biometric future
By Sam Williams
Ten years ago, Dr. Joseph Atick was a Rockefeller University research scientist sitting atop an intriguing and potentially lucrative breakthrough in the realm of pattern recognition.
In an attempt to mimic how the human brain processes sensory signals, Atick and his research team developed a computational model that zeroes in on a person's facial landmarks and measures the relative distances between them. Once stored, these measurements become, in essence, a template unique enough to match individuals with their photo-ID or mugshot images in various, controlled situations.
For Atick, a mathematical physicist by training, the future boiled down to two choices: He could debate the implications of that breakthrough amid the safe, quasi-utopian world of academia, or he could try to put it to work in the messy, dynamic world of commercial software.
"It was almost like a curse," says Atick, now the chief executive officer of Identix, a Minnesota-based leader in fingerprint- and facial-recognition technology. "You're cursed with the blessing of knowing something important, something that society wants. You feel like it belongs to society and not to you."
The notion of curses attached to certain elements of human knowledge is an apt introduction to the field of biometric software, the blanket term used to describe software built to identify or authenticate human users from digitally captured physical data. Like genetic engineering and nuclear power, it is a field where innovation and controversy go hand in hand and where ordinary researchers must constantly weigh the long-term implications of their work. Toss in the vagaries of a marketplace where governments, police departments and other slow-moving bureaucracies play the role of lead customers and early adopters, and the "curse" image seems almost too good to ignore.
As the field's own evangelists note, it is an industry niche where optimism and reality often have a hard time making a connection.
"The perception out there is that with all this money the government is funneling into [biometric security], it should be growing like gangbusters," says Jeff Watkins, a senior technology consultant for the International Biometric Group, a New York consulting firm. "Although you do have a number of companies that are growing, they're certainly not growing with the speed that we saw with other software and technology segments."
Watkins' own company has put the sector's rate of growth at 44 percent per year, a rate that would put the global biometric industry over the billion-dollar mark by the end of the 2004. Such numbers date back to 2002, however, when the sector was enjoying a sustained period of positive press in the wake of the Sept. 11, 2001, terrorist attacks.
Since then, the media's attention has chilled slightly as journalists have zeroed in on the ethical and technological challenges that still impede the use of biometric security as a cheap replacement for flesh-and-blood security. As civil rights groups have focused their ire on the U.S. PATRIOT Act and other post-9/11 expansions of federal policing authority, biometric companies have, for the most part, retreated into the background of the dispute.
For biometric entrepreneurs, the dimming of the media spotlight has brought an associated decrease in venture capital and IPO prospects. Still, given the controversial nature of the business, many seem to have welcomed the recent quiet.
"After 9/11, a lot of noise entered the system," says Atick. "In a way it ended up delaying things, because people got confused about what was available. Luckily the confusion goes away through a natural-selection process."
The quiet period may be ending. Last month, the U.S. Department of Homeland Security unveiled its controversial U.S. Visitor and Immigrant Status Indicator Technology, or US-VISIT program. Budgeted at just over $1 billion for the next three fiscal years, the program will install electronic fingerprint scanners and cameras at 115 U.S. airports and 14 U.S. seaports. At the moment, only visitors from certain countries must submit to fingerprinting. Exempt countries have a limited period to implement their own face-recognition or machine-readable passport systems in accordance with recommendations put out by the International Civil Aviation Organization, or ICAO, the United Nations standard-setting body for international air travel.
The American Civil Liberties Union has been quick to criticize the US-VISIT program. Jay Stanley, spokesperson for the ACLU's Technology and Liberty program, says US-VISIT epitomizes the uphill fight privacy advocates have to face every time a large-scale biometric system is rolled out.
"You have to look at the incentives," Stanley says. "For companies and governments, the incentives associated with biometrics all point the other way from privacy. The incentives are for more data collection and better linking of different records until, finally, you have the ability to create a rich portrait of a person's life and activities."
Biometric entrepreneurs don't entirely disagree. Andy Amanovich, senior technology strategist for and co-founder of Imagis Technologies, a Vancouver facial-recognition software firm, says a sizable portion of his job is devoted to dampening expectations on the customer side.
"I think there really is a hope out there that you can take a database with the names of 1,000 al-Qaida terrorists, wipe off the sand, and start matching those names to faces as they pass through an airport," says Andy Amanovich.
That his company is unable to offer such capabilities is encouraging from both a technical and an ethical perspective, says Amanovich. Clandestine surveillance systems that can capture a facial image or iris scan off a moving target and compare that image to an existing database may be common in movies like "Mission Impossible" and "Minority Report," but in reality, they are a few decades away in terms of mathematics and processing power.
"Which I'm glad about," says Amanovich. "I happen to like freedom and democracy."
Amanovich's own company exemplifies the field's current, unhyped capabilities. Imagis sells a facial-recognition software package dubbed ID2000 to police departments for use in patrol car computers. During an arrest, police take a digital photo of the suspect and run it against an existing database of mug shots.
As in most other software systems, the really heavy lifting is conducted on the back end, and Imagis supplements its offering with a Microsoft .Net-based set of services and tools designed to facilitate data sharing across a distributed, Internet-based system.
Although there are still privacy implications, the use of ID2000 builds on existing methods of police procedure and adheres to constitutional safeguards such as probable cause, which makes it less of a red flag for groups such as the ACLU.
"I don't think we'd have as much of a problem with something like that," says Stanley. "Our biggest problems with face recognition come in situations where it's used for routine surveillance of individuals and from the fact that it's ineffective as a security measure."
For a look at systems that draw the full ire of groups like the ACLU, one need only go back a few years to FaceIt, a closed-circuit facial-recognition system used by the Tampa Police Department in the Ybor City nightlife and entertainment district from 2001 until last year. Built by Visionics, a pre-merger incarnation of Atick's Identix, the system was doubly cursed. Technical glitches limited its crime-fighting performance to the point that the Tampa police were unable to claim a single positive ID after two years of performance. The negative publicity, meanwhile, helped forge an unlikely alliance between the ACLU and prominent conservatives, including Republican Rep. Dick Armey, then the House majority leader.
"We are taking a step in the wrong direction if we allow this powerful technology to be turned against citizens who have done no wrong," said Armey at the time.
Two years later, Atick seems unscarred by the criticism aimed at FaceIt. For one thing, the events of Sept. 11, 2001, have reduced the general public's aversion to biometric security. Second, Atick sees public disapproval as a helpful part of the technology-shaping process. Viewing the marketplace in Darwinian terms, Atick sees outrage as the functional equivalent of natural selection.
"Society, by defining which technologies are going to be adopted, can decide which technologies are viable," Atick says.
Noting the current market niches, Atick says the alarm stirred up by the 2001-2003 FaceIt trial has reinforced the need for a stronger display of the quid pro quo bargain inherent in most security systems.
"When you capture a biometric you have to be sure that you as the end user own the biometric," Atick says. "The reason I'm giving it to you is because you're giving me something, some privilege in exchange. As long as you give me something in exchange and I agree to give you this, and there are groups to audit that process, I think we can create a situation where people can use the technology without too much hesitation."
Still, given the evolving dynamics of the field and lingering misconceptions about the current abilities of most biometric systems, some entrepreneurs see a need to speed the decision-making loop. Serge Belongie, co-founder of the fingerprint-recognition company Digital Persona, is a UC-San Diego computer scientist who, unlike Atick, chose to remain in academia after releasing his technology innovations into the wild. The decision has not distanced him from the ethical implications of the field, however, so in 2000, doing double duty as Digital Persona's chief research officer, Belongie helped put together an exhibit on biometrics and body scanning at the San Francisco Exploratorium.
Titled "Revealing Bodies," the exhibit was a five-kiosk display. At the first kiosk, Belongie says, visitors registered their fingerprints and entered additional personal information. At the next three kiosks, visitors who submitted to fingerprint scans were welcomed back, by name, and given additional information about other, related technologies. At the final kiosk, users received a notice that their fingerprint templates and responses to queries would be deleted from the system but not immediately.
"Basically, we said, 'Now think about what happened. We know all these things about you. How does that make you feel?'"
In the resulting feedback forms, emotions ranged from awe to anger, Belongie says. Though some protested the fingerprint scans, many offered questions like "What happens if I lose my finger?" "What if somebody fakes my fingerprint?" or "How can I get one of these myself?" Overall, Belongie says, respondents were "very engaged" and optimistic about the potential uses.
"They liked that a biometrics company was asking these questions," he says.
Looking at alternative uses like US-VISIT, Belongie, whose own company has shied away from the government-services market, has yet to make up his mind on whether such uses deserve the same optimism.
"Way back in 1993 when we started doing this stuff, I had no idea if it would succeed, but I could imagine people putting down a thumbprint to get on an airplane," Belongie says. "That said, I have not made up my mind yet about the pros and cons of US-VISIT. As a university professor, the so-called war on terror concerns me very much because of the difficulties it is causing for international students. I can see how performing biometric recognition at airports and seaports could catch known terrorists who are attempting to arrive in the U.S. via airports and seaports. This possible benefit provides me little comfort, however, in the face of escalating hatred and resentment toward the U.S."
Atick, on the other hand, considers a US-VISIT a "coup" both for the industry and for the U.S. government which, in citing the ICAO call for biometric security safeguards, has put other developed nations in a tough position to condemn the move. That said, Atick, noting the curse that brought him into the field, says the resulting political backlash, if it ever emerges, will be equally helpful for setting the future course of the industry.
"This is not an industry you can be neutral about," Atick says. "This is an industry that changes the way people travel and provide information to one another. If you're out of touch with people's attitudes about it, you're going to face significant resistance. If on the other hand you're going to provide a beneficial service, you're better aligning yourself with where society is heading."
- - - - - - - - - - - -
About the writer
Sam Williams is a freelance reporter who covers software and software-development culture. He is also the author of "Free as in Freedom: Richard Stallman's Crusade for Free Software."