DHS looking into body odor-reading devices

The Department of Homeland Security plans to study the possibility that human body odor might be used to determine when people are lying, or to identify individuals in the same way that fingerprints can. In a federal procurement document posted Friday on the Web, the department´s Science and Technology Directorate says it will conduct an "outsourced, proof-of-principle study to determine if human odor signatures can serve as an indicator of deception. ... As a secondary goal, this study will examine ... human odor samples for evidence to support the theory that an individual can be identified by that individual´s odor signature." Officials said the work was at a very early stage, but the announcement brought criticism. Barry Steinhardt, of the American Civil Liberties Union's Technology and Liberty project, said the plan showed that the department had "misplaced priorities." "The history of DHS´ deployment of these technologies has been one colossal failure after another. There is no lie detector. This research has been a long, meandering journey which has taken us down one blind alley after another." Mr. Steinhardt said that even well-established biometric identity technologies such as fingerprinting had resulted in inaccurate identifications of people, including Oregon lawyer Brandon Mayfield, who received an apology from the FBI after being wrongfully accused of having had a hand in the 2004 Madrid train bombings. "None of the biometrics for identity have worked very well, with the possible exception of DNA," he said, adding that even fingerprint evidence was "increasingly being challenged in courts around the country." The procurement notice says the department is already "conducting experiments in deceptive behavior and collecting human odor samples." The research it hopes to fund "will consist primarily of the analysis and study of the human odor samples collected to determine if a deception indicator can be found." "This research has the potential for enhancing our ability to detect individuals with harmful intent," the notice says. "A positive result from this proof-of-principle study would provide evidence that human odor is a useful indicator for certain human behaviors and, in addition, that it may be used as a biometric identifier." Amy Kudwa, a Homeland Security spokeswoman, said "proof-of-concept" work was the earliest stage of technological development. The directorate "is trying to determine what factors of human behavior and chemistry can provide clues to the intent to deceive," she said. The work would be carried out by the Federally Funded Research and Development Center run by the nonprofit Mitre Corp., which conducts cutting-edge research for the U.S. military, Homeland Security and intelligence agencies, Ms. Kudwa said. Scientific research shows that so-called volatile organic compounds present in human sweat, saliva and urine can be analyzed using a technique known as gas chromatograph-mass spectrometry. Research published by the Royal Society in London in 2006 found "a substantial number of marker compounds [in human sweat] that can potentially differentiate individuals or groups." Researchers took five samples each from 179 people over a 10-week period and analyzed them, finding hundreds of chemical markers that remained more or less constant for each person over time. An analysis of these compounds "found strong evidence for individual [odor] fingerprints" the researchers concluded. They warned, however, that some people appear to have less distinctive odors than others. "The reason for the variation in distinctiveness is unclear," the researchers said. More importantly, the odors of some people changed during the course of the study. "Not all subjects had consistent marker compounds over time, which might be due to physiological, dietary, or other changes," they concluded. The researchers also cautioned that some of these marker compounds might be "exogenous chemical contaminants," from skin care or perfume products, or tobacco smoke and other substances present in a person's environment. About a quarter of the 44 apparently distinctive marker compounds they were able to analyze appeared to be artificial contaminants, the researchers said. "Determining the origins of individual and sex-specific odors - and controlling exogenous chemical contaminants - may provide the most important challenge for future ... studies," the researchers said. Analysts said those challenges are likely to be significant, and they will multiply if the techniques are deployed in the field. "While some of these sensors perform well in the lab, the real world may be different," said technology consultant and author John Vacca. "The technology is still in its infancy."