Can Robots Better Spot Terrorists at Airports?
By Jack Nicas, WSJ, Dec. 30, 2013
Next to have their jobs automated: airport-security screeners?
Aviation and government authorities are starting to use machines in lieu of people to verify the identities of fliers by scanning their faces, irises or fingerprints. Dozens of airports in Europe, Australia and the U.S. already employ such technology so passengers can pass immigration checks without showing identification to, or talking with, a person. Now, several major airports in Europe have started using these automated ID checks at security checkpoints and boarding gates.
The use of biometrics—computers verifying identities through physical characteristics—and other automated techniques in airport security is raising questions about the strengths of man versus machine in detecting potential terrorists. Industry officials argue the advantages outweigh the risks, and are promoting automation to help make air travel more efficient and less frustrating—and to save money.
Ultimately, the technology could “get rid of the boarding pass completely,” with fliers’ faces serving as their tickets, said Michael Ibbitson, chief information officer of London Gatwick Airport. Gatwick performed a trial this year in which it processed 3,000 British Airways fliers without boarding passes. The fliers scanned their irises when checking in, enabling cameras at security checkpoints and boarding gates to automatically recognize them. “We’re only just starting to see what biometrics can do,” he said.
Proponents, including government and industry officials, say that automation of airport security holds the promise to free human screeners so they can focus on detecting suspicious behavior or monitoring flagged travelers. And for some aspects of security, they say, computers can be more thorough and less error-prone than humans.
Critics, however, worry that relying too much on automation will dull the senses of human screeners and remove the human intuition that can detect when something just doesn’t seem right.
About 28% of the world’s airports now use biometric technology, up from 18% in 2008, according to a survey by SITA, an airline IT provider.
The International Air Transport Association and Airports Council International, two of the industry’s largest global trade groups, advocate automation as part of their initiative to streamline airport security. The groups say the lengthy and cumbersome security process is deterring some travelers from flying, and note that the average checkpoint now processes about 150 passengers an hour, half the rate before the terror attacks of Sept. 11, 2001.
Their “Smart Security” program will use new checkpoints at Amsterdam Airport Schiphol and London Heathrow Airport next year to test key proposals, likely including new baggage-screening software that automatically clears some simple objects like clothing without running X-ray images past human screeners.
Guido Peetermans, the program’s manager, said explosives are typically very dense, but low-density items like jackets can be automatically cleared. “The reliability of the algorithm is very high,” he said. “Technically, what it means is there are a certain number of images that no human operator will see.”
Airlines, airports and authorities all say that aviation security will remain a layered process that always includes some human interaction, intelligence and randomization. Major security changes, particularly at checkpoints overseen by government agencies, will take years to come to fruition.
The U.S. does use biometrics for airport border control. Nearly 2 million frequent fliers are enrolled in trusted-traveler programs under the U.S. Customs and Border Protection that let them scan their fingerprints instead of talking to an immigration officer when re-entering the country. These biometric kiosks were used 820,000 times this summer, up 75% from the summer of 2012.
The U.S. also collects photographs and fingerprints from every foreign national entering the country.
In a test of facial-recognition scanners at Amsterdam Schiphol in 2010, Dutch university and government researchers found that the machines could correctly approve almost 98% of travelers, while allowing one of every 1,000 impostors to pass, on average. The study didn’t give comparable numbers for humans.
Still, in 2011, Gatwick’s automated immigration kiosks twice approved travelers who were using the wrong passport, and the kiosks at the Manchester, England, airport were temporarily closed because of a similar incident, according to the U.K.’s chief immigration inspector.
Richard Bloom, director of terrorism, intelligence and security studies at Embry-Riddle Aeronautical University in Prescott, Ariz., said computers may be more reliable than humans, but they are also more predictable—making them vulnerable to sophisticated terrorists. “As long as you can find out how a machine works,” he said, “people with resources can figure out how to beat it.”