Is Predictive Profiling Legal?
This is a guest post from Jason Riley.
One question that may be on the mind of law enforcement personnel is, "Is combat profiling (or predictive profiling) legal? This is a topic which I would like to address briefly in this post.
As PVH has discussed in a previous post, combat profiling is basically an application of predictive profiling. We give it particular terminology, and form applications which relate directly to combat and law enforcement situations. In regard to the legality of observing behavior to identify potential threats, I would like to address a few issues. The basis for this discussion is a paper written by two Non-Resident Fellows of the Georgetown Center on National Security and the Law, which addresses this very question (the article can be found here: http://www.georgemasonlawreview.org/doc/17-2_FlorenceandFriedman.pdf).
The first thing that I should discuss is what we mean by threat. While this could be defined in a number of ways, I often speak of "threat" in the sense of someone who intends to do another person harm. We do not limit our observation to this particular type of person, but a person seeking to do another person harm is certainly at the top of the list. However, we are also looking for other "anomalies;" people conducting surveillance, observers, trigger men (military personnel will know what this term means), and anyone else with less than good intentions. For a cop this could mean (in addition to people who want to do them harm) people casing a joint, a predator looking for victims for his next crime, a fleeing perpetrator, the girl who just stole something, etc. But back to the question at hand. Is observing behavior for certain indicators and making proactive decisions based on those indicators a legal way to identify a potential threat (or other persons of interest)?
This is certainly a legitimate question, particularly in our rights-conscious society, and especially in light of the increased security measures being introduced in places like airports. Some of these security measures are arguably intrusive, and may actually push the envelope in regard to violating a person's constitutional rights. But does predictive profiling violate a person's constitutional rights?
In an article discussing the legality of similar initiatives by government agencies (such as the SPOT program), Justin Florence and Robert Friedman, ask the following questions:
"But these mass behavioral profiling programs raise serious legal and constitutional questions. Are they really just covers for racial profiling? Do they violate the Constitution’s equal protection principles? The Fourth Amendment’s prohibition on unreasonable searches and seizures? Are they fundamentally unfair? Can they easily be abused? How can the government be kept in check if the public does not even know what agents are looking for? Most importantly, given the secrecy required to operate these sorts of national security programs, what oversight institutions can keep them in line with our constitutional commitments?" (Florence and Friedman, "Profiles in Terror," p. 424)
We have already discussed that combat profiling is not racial profiling. We are not concerned with skin tone, ethnicity, or religion. We profile based on behavior. In regard to their discussion of the legality of behavioral profiling, Florence and Friedman discuss behavior profiling in regard to three aspects of the constitution: 1) the Equal Protection Clause (Amendment XIV), 2) the Fourth Amendment right against unlawful searches and seizures, and 3) the Due Process Clause (Amendment V). After a lengthy legal analysis, the authors conclude that at least the SPOT program, which is fundamentally similar to our method, "appears to be consistent with these three provisions of the Constitution."
Interestingly, the authors do provide a few criteria that a program such as this should meet to ensure that it does not run afoul of the Constitution (Florence and Friedman, "Profiles in Terror," p. 463). I will list the criteria in bold and then discuss it from our point of view.
1) The program’s design is based on valid, scientific theories supported
by actual evidence. Our principles are based on scientific studies and are deeply embedded in theoretical and applied scientific fields. We based our methods and principles on the work conducted in kinesics, proxemics, heuristics, etc. We consider research done by individuals such as Sandy Pentland (MIT) and Paul Ekman (pioneer and expert in micro-facial expressions).
2) The scientific theories translate successfully from a laboratory environment
to an actual, real-world security operation. The Israelis, American law enforcement agencies, and deployed Marines use techniques such as this with success.
3) The program, in design and actual effect, avoids racial, ethnic, or religious
profiling. We continually repeat that race, ethnicity, and religious background are not criteria we use to determine who is and who is not a threat. The fact of the matter is that every person we observe is going to have a skin tone, ethnicity, and religion (or lack thereof), but what we focus on is behavior.
4) The security agents responsible for implementing the program are sufficiently trained and are able to implement the program correctly and consistently in practice, and do not bring in any racial, ethnic, or religious bias––consciously or not. Unfortunately, this criteria is not something that can be evaluated. What level of training is considered "sufficient"? A day, a week, a month? Additionally, other scholarly fields have shown that there is no way to absolutely get rid of bias. So, I would revise this criteria to say that the program, and the instructors, provide explicit guidance to prevent bias, and supervise the conduct of security agents to ensure that bias is as limited as possible.
5) The program is not abused or used for inappropriate purposes, such as harassing individuals. We intend that our principles are used to protect oneself and others. Our main purpose is safety and security. None of our methods should be used to harass or manipulate anyone.
6) The individuals affected by the program are not repeatedly wrongly targeted, and, if they are, they have some opportunity for redress. I will address this criteria in two ways. First, since we focus on behavior, it is possible that certain individuals may be identified as an anomaly more than once, IF their behavior consistently provides indicators that they may pose a threat. Second, since we are also concerned with establishing a baseline, a certain anomalous behavior by a particular person may become a baseline behavior. Once this is established, that person should not be identified as an anomaly unless a different factor comes into play.
7) The program is updated to respond to and reflect new developments in science and world affairs, so that it can minimize the targeting and inconveniencing of innocent travelers. We are constantly reading, researching, and training to advance our understanding of behavioral/predictive profiling to include increasing our knowledge of human behavior, psychology, and related topics.
8 ) The program provides effective security. For this criteria, the "proof is in the pudding." We consistently receive feedback that the techniques, methods, and principles we teach keep people alive and help them make the best decision in uncertain environment.
From our perspective, combat profiling stacks up pretty well against these criteria. Marines, soldier, police officers, and other law enforcement personnel should use these principles and skills with confidence that they are not violating anyone’s rights and are observing the most accurate indicators of potential threats: behavior.
Let me end with a lengthy quote from their article:
"Since 9/11, public debate on the appropriate domestic response to the terrorist threat has focused on balancing civil liberties and the rule of law against the need for national security. While both critics and proponents of
behavioral profiling programs may wish to frame the debate through this narrative, we believe this approach is unnecessary and unhelpful. For the reasons we described in Part I, the behavioral profiling paradigm makes a good deal of sense. It allows security officials to identify and focus resources on the people who are most likely to be a security risk, with relatively little cost. Nonetheless, as we acknowledge in Part II, behavioral profiling raises serious constitutional concerns under the Equal Protection Clause, the Fourth Amendment, and the Due Process Clause. The security benefits of behavioral profiling need not be balanced against these constitutional and civil liberties concerns. For, as a matter of doctrine, the fundamental issue is whether behavioral profiling programs are reasonable: whether they provide a rational basis for distinguishing between some persons and others, whether they provide reasonable suspicion for a law enforcement officer to conduct a follow-up search and seizure, and whether they treat people fairly, including by offering some redress for wrongly harmed persons" (Florence and Friedman, "Profiles in Terror," p. 481, emphasis mine).
We believe that combat profiling aligns well with the comments above. First, it makes a good deal of sense. Second, the program is reasonable. It provides security personnel a rational basis for making decisions to determine reasonable suspicion to take action on a perceived potential threat.
Questions, comments? Let me know.