By Jay Stanley
Back to main ATS Page
Recently, it was revealed that the government has improperly diverted a tool for screening cargo, the controversial Automated Targeting System (ATS) program, into a data mining program that assigns all who cross the nation's borders, citizen and non-citizen alike, with a computer-generated "risk assessment" score that will be retained for 40 years.
Many fundamental questions remain unanswered about ATS. The program raises concerns because it brings innocent Americans under government suspicion without justification and it threatens our privacy while undermining our constitutional freedoms. Congress must investigate this misappropriation of scarce homeland security dollars and congressional intention. Congress must also reassert the American values of limited intrusion into Americans' private lives.
Conflicting statements about a mystery program
The Administration has issued contradictory statements about ATS. For example, Stewart Baker, the Assistant Secretary for Policy at the Department of Homeland Security, claimed in a recent speech that ATS does not involve the scoring of human beings.1 But Baker's assertion is flatly contradicted by other official statements by DHS, including a legally binding Privacy Act "System of Records Notice" (SORN) on this program published in the Federal Register ("ATS builds a risk assessment for cargo, conveyances, and travelers based on criteria and rules developed by CBP"2) as well as an ATS "Privacy Impact Assessment" ("Every traveler and all shipments are processed through ATS, and are subject to a real-time rule based evaluation. ATS provides equitable treatment for all individuals in developing any individual's risk assessment score."3
The need to get real answers
Such contradictions highlight how little the public truly knows about ATS, and reaffirm the need for Congress to investigate so that it can be publicly debated. Many fundamental questions remain unanswered about ATS. Is there a risk assessment score? Is it in effect now? Is it retained? If so, how is it used? This subject is too sensitive for policy to be formulated on the ground and in secret by invisible bureaucrats and security agents.
Congress must also investigate the powers Homeland Security is claiming for itself for potential future use, without statutory authorization to substantiate that claim. For example, even if the agency is using its discretion not to maintain data for 40 years, as officials have protested is most often the case, the fact remains that the agency has reserved for itself the authority to do just that anytime it pleases.
Vague and misleading claims about the program's successes
DHS officials also advanced broad claims about the purported successes of the ATS program. In Baker's speech, for example, he cited specific anecdotes in which individuals were captured (a man who later carried out a suicide attack in Iraq, a man found with bomb-building literature, and a woman who was later discovered to be working for a human-smuggling program) as a way of dramatizing the program's purported successes.
We can understand why DHS would want to give out to the public these kinds of stories to support their program. But since we do not really know exactly what ATS is or how it works, the public has no idea how these successes were achieved, and what role if any ATS played in them. More specifically, the public has no ability to judge whether computerized data searches and profile-building (the controversial activities in this program) played in them. Unanswered questions include:
Were these catches the result of pure luck? Tens of millions of individuals enter the United States each month, and our customs officers have always turned away some from our borders. Given the millions of people who cross our borders, people are turned away for many reasons in the course of regular business. For example, their papers may not be in order. Yet, these people may have no connection to terror or crime. If Customs and Border Protection ("CBP") were using psychics or palm-readers to pick travelers for further scrutiny, pure chance would still guarantee the agency some success stories it could point to in a speech. But such isolated cases tell us nothing about whether this system is truly effective.
Does the government even know whether the system is effective? Baker makes no mention of any systematic audit or other evaluation by the government to determine whether this program is more effective than blind chance, and enough so to justify its costs. For example, it is unknown whether the government simply assumes that every person turned away represents a success - that such persons were in fact guilty of whatever suspicion led to their rejection.
Were these catches the result of old-fashioned sleuthing? In one purported success cited by Baker, an individual was denied entry based upon an interview by a customs officer, and the fact that he could not withstand such in-person human scrutiny. Such interviews long predated the invention of computerized "risk assessments," and Baker provides no reason to believe that computer risk scores contributed at all to the decision not to admit him. In another anecdote, Baker claims that ATS allowed agents to take down an international human-smuggling ring. But the destruction of that ring was a result of old-fashioned, predicate-based investigation, and in no way justifies the collection and maintenance of data on millions of Americans and the construction of computer-generated "threat scores" on them, which is the activity that is controversial in this program.
Were these catches the result of simple watch list checks? Watch lists are a primary part of ATS, and Baker says simply that "data in the DHS computer system flagged" these individuals. If watch list comparisons led to Baker's claimed "successes," then programs other than ATS deserve the credit and ATS risk score assessment of innocent travelers remains unjustified.
A dangerous road to follow
Although government spokespersons strain mightily to portray ATS as a harmless, commonsense system, it is just the kind of data mining approach to terrorism that raises very serious and fundamental questions about privacy.
Current limits are not likely to stand
There is, in fact, every reason to believe that this program will become more sweeping. Just how much information about a person's background does the government need in order to truly establish that someone is not a threat in any way?
There are no obvious natural limits to the data in which the government might find "patterns that establish reason to ask questions," as Baker put it. Does the government get to scrutinize every address at which you've ever lived? Quiz you about the fact that you once went for 12 months without a job? About your Web surfing, your online book purchases, your school transcripts, your internal movements, your associations, your ex-spouse?
That is all data that could be conceivably be "relevant" to an attempt to detect a terrorist threat, or criminal wrongdoing. That is all data that could become part of the ATS system within the "limits" defined by the notice published in the Federal Register. If the government really intends to use information only in much more limited ways, then such limits should be enacted into hard law.
Because you can't prove a negative, it is entirely predictable and indeed inevitable that over time, if not checked by an insistence on our collective values of privacy, the government will push for more and more sources of information out of which to build these ratings.
To the question of how much data the government should have access to, the implicit answer given by ATS's defenders is "as much as it needs." But that "need" will prove bottomless unless it is counterbalanced by our ideals of privacy and limited government - balancing values that are missing from the design of this program. Perhaps it is too much to expect that those who work in our security agencies can themselves keep in perspective those larger counterbalancing American civic ideals. For that job, the American people must rely upon Congress.
(In fact, Congress has already acted several times to do just that, having included an explicit ban on the risk scoring of innocent individuals in the Homeland Security appropriations bills for 2005, 2006 and 2007. However, with the ATS program, DHS directly ignored that statutory prohibition. The agency's disrespect for clear statutory law and congressional intention cannot go unnoticed .4)
DHS presents the public a false choice
Americans are not faced with a choice between allowing government background checks without limits, and leaving themselves open to terrorism. As many security experts have pointed out, the most effective way to stop terrorism - and the means by which most real terrorist plots such as the London liquid bomb plot are actually thwarted in real life - is old-fashioned investigatory work based on evidence and leads, rather than dragnets that treat everyone as suspect. The problem with systems like ATS is that they collect a lot of information that is not helpful. It does not differentiate between evidence and data. As a result, instead of focusing our attention on the greatest threats, ATS directs scarce homeland security resources to wrongly casting a net of suspicion over millions of innocent Americans.
Finally, ATS subverts the Fourth Amendment by allowing DHS to create a dossier on every American Traveler. In short, this program turns every American traveler into a criminal and terrorist suspect. No one objects to the keeping of legitimate law enforcement files - such as information about the suspects and perpetrators of the World Trade Center attacks or any other crimes - for extended periods of time. But why does the government need to grant itself the authority to maintain risk scores or other information on every traveler for 40 years? The only apparent reason that policy would make sense is if the government wants the ability to look back at any person's life for 40 years just in case they become a suspect, or in case the old information is later deemed useful for some future computerized sweeps. But in this country, the government is not supposed to collect information on its citizens just in case they commit a crime.
1 "Risk score - where did this come from? . . . We have risk scores for Cargo. . . but I don't think that we're scoring human beings, and we're certainly not keeping scores on them, so there's nothing to disclose there." Baker, remarks at the Center for Strategic and International Studies, Washington, D.C. December 19, 2006; transcript online at http://www.csis.org/media/csis/events/061219_baker.pdf.
2 "Notice of Privacy Act system of records." Federal Register, Volume 71, Number 212 (2 November 2006): 64543-64546; online at http://edocket.access.gpo.gov/2006/06-9026.htm. Emphasis added.
3 "Privacy Impact Assessment for the Automated Targeting System," Department of Homeland Security, Nov. 22, 2006; online at http://www.dhs.gov/xlibrary/assets/privacy/privacy_pia_cbp_ats.pdf. Emphasis added.
4 DHS's statutory interpretation challenging this argument is flawed. DHS officials argue that the ban applies only to the Secure Flight program. But, although the provision in question (514(e)) is adjacent to provisions that apply only to Secure Flight (and which explicitly state as much), the language banning risk algorithms makes no mention of Secure Flight, and instead is a catch-all that says "none of the funds provided in this or previous appropriations Acts" may be used in that way.