In 1985, Alex J. Jeffries and his colleagues demonstrated that patterns of molecular markers in human DNA, or DNA fingerprints, could serve as uniquely identifying personal traits. This discovery was quickly applied by the criminal justice system, as way of definitively connecting suspects with blood, tissue, or semen from crime scenes. Shortly thereafter, governments at the state and national levels began authorizing the collection of DNA samples from individuals convicted of violent crimes who were considered at high risk for recidivism. By 1998, all fifty states in the United States had enacted such laws, and the U.S. Federal Bureau of Investigation (FBI) was able to launch a national electronic database of DNA profiles from convicted criminals for use in future cases (Hoyle). In the interim, the collection of DNA for personal identification purposes has already become mandatory within the military and has become a mainstay of civilian efforts to clarify the identities of children and kidnap victims, to investigate family lineages, and even to authenticate religious relics. On the horizon, lies the question that civil libertarians anticipate with dread: Why not store personally identifying genetic information on everyone as a matter of course, for the advances in public safety and personal security that can be gained thereby?
Photographs and traditional fingerprints have, of course, also been taken, collected, and used for all these same purposes in the past. But unlike photography and manual fingerprinting, collecting individually identifying DNA patterns (iDNAfication) does involve taking bits of people's bodies from them: nucleated cells and their complements of DNA molecules. For those concerned about the ethical and legal status of body tissues and an individual's ability to control what happens to him or her through use of that tissue, this corporeal side of iDNAfication raises an interesting challenge. Clearly, questions of personal privacy are involved. But unlike most other disputes over body tissues, the issues here are not primarily matters of personal sovereignty.
For example, unlike involuntary sterilization or forced surgeries, the central concern with mandatory iDNAfication does not seem to be the violation of a person's bodily integrity. Compared with the other infringements of personal freedom that legitimately accompany legal arrest, providing a saliva or cheek swab sample seems negligibly invasive (Schultz). Moreover, unlike the creation of marketable human cell lines or the commercialization of organ procurement, it is not the exploitation or misappropriation of the person's body for others's gain that is centrally troubling either. Manual fingerprints and photographs also exploit suspects's bodies in order to incriminate them, without raising special privacy concerns. Moreover, consider the fact that it does not matter to an identical twin whether a DNA sample under scrutiny actually comes from him or his sibling: To the extent that the genetic information it contains describes both their bodies, the privacy of each is endangered.
In fact, the major moral concern about iDNAfication has little to do with whether the DNA analyzed is a piece of the person being identified, the property of the person being identified, or even is forcibly extracted from the person being identified. In most iDNAfication contexts, these physical, proprietary, and decisional privacy considerations are beside the point. Rather, the important feature of iDNAfication is what the DNA analyzed can disclose about the person being identified. It is, in other words, individuals's informational privacy that is at stake in the prospect of widespread iDNAfication, and it is in those terms that the policy challenge of iDNAfication should be framed. What should society be allowed to learn about its citizens in the course of attempting to identify them?
Taking up this challenge means taking seriously the precedents set by society's use of photography and manual fingerprinting, since their primary impact on personal privacy also lies in the identifying information they record rather than the nature of their acquisition. If the collection of mandatory mug shots and fingerprint impressions are taken as benchmarks of social acceptance for at least some identification purposes, any iDNAfication methods that conveyed no more personal information than those techniques should also be socially acceptable, for at least the same range of purposes. Thus, where fingerprints of arrestees, inmates, employees and recruits are now taken legitimately, performing iDNAfication should also be justified, if its informational privacy risks were equivalent. Similarly, if society accepts the personal disclosures involved in using photographs on drivers's licenses and identification cards, it should be willing, in theory, to expose an equivalent range of genetic information in any legitimate forms of iDNAfication. One approach to the general challenge of iDNAfication, then, would be to ask the following question: If the ways in which photographs and manual fingerprints are used for legitimate identification purposes are accepted, under what circumstances, if any, might forms of iDNAfication meet the standard those practices set for the disclosure of personal information?
Personal Privacy Considerations
A number of personal privacy risks of iDNAfication have been described and anticipated in the design of some iDNAfication programs. Thus, for example, many have pointed out that if the DNA sequences used as the components of an iDNAfication profile are taken from the regions of the human genome that code for proteins, important biological information about their sources could be revealed, including information about their paternity, current health status, and potential health risks (U.S. Congress Office of Technology Assessment (OTA), National Academy of Sciences). Any risk of disclosing sensitive personal information of these sorts would clearly increase the intrusiveness of iDNAfication beyond that of traditional fingerprinting and photography. In addition, it could expose the person being described to the possibility of discrimination on the basis of a disclosed genotype (Bereano; DeGorgey; Scheck; Sankar) Fortunately, this is a privacy risk that can be almost entirely eliminated by two simple precautions: One need only avoid analyzing biologically informative DNA, and destroy the DNA samples upon analysis.
The first precaution can be accomplished by restricting the sections of DNA that are amplified, analyzed and utilized in the iDNAfication profile to the non-coding regions of DNA between our functional genes. By definition, markers selected from these regions will not disclose any biologically significant information. Rather, like fingerprints, they could merely provide a unique pattern to match in seeking to identify an unknown person. Even photographs are useful mainly as patterns to match, rather than for what they can independently tell us about the person pictured in them. Serendipitously, individual variation is also vastly more pronounced in this so called junk DNA (since mutations can accumulate in these sections without having any adverse effect on genomic function), making it more attractive for iDNAfication purposes on scientific grounds as well.
Thus, the FBI, in establishing standardized forensic iDNAfication markers for use by state laboratories contributing DNA profiles to the latter's National DNA Index System (NDIS), has focused on a set of thirteen loci from non-coding regions that contain series of repeated nucleotide sequences whose length is highly variable between individuals (Hoyle). The exclusive use of these markers in any iDNAfication program would forestall most genetic privacy concerns linked to the biological information content of the DNA profile itself.
The second important step to insuring the genetic privacy of iDNAfication is to destroy the physical samples of DNA once DNA profiles have been generated from them. As long as the DNA samples themselves are retained, the risk remains that they could be retested for their biological informational content. Thus, in its report on forensic DNA analysis, the National Academy of Sciences in 1990 recommended that even samples taken from convicted offenders be destroyed promptly upon analysis, and the FBI has designed its national iDNAfication collection as a databank, not a DNA bank, including only the electronic profiles of non-coding DNA markers (Murch and Budowle).
This second precaution has not been adopted by forensic laboratories at the state level, or by the military at the federal level. Most of these laboratories plan to bank their actual DNA samples indefinitely, on the grounds that the samples may need to be retested as new markers or testing technologies become standard (McEwen). The Department of Defense is storing dried blood samples from its recruits, for genotyping only in the event that the recruits later turn up missing in combat. This effectively undercuts the privacy protections afforded by using non-coding markers in the iDNAfication profile itself, and immediately elevates the privacy risk of any iDNAfication program well beyond that of ordinary fingerprinting. Even if, contra the National Academy of Sciences, this increased risk were tolerable for convicted offenders, it should not be for military recruits, government employees, or arrestees, since the potential intrusion goes well beyond what is required for identification.
Social Policy Considerations
Despite the initial hopes of early enthusiasts like English scientist Francis Galton (1822–1911), large collections of ordinary fingerprints have never been useful for much else besides individual identification. (Rabinow) The informational potential of the human genome, however, does require the designers of iDNAfication systems to consider in advance the range of uses they should accommodate. Even when a DNA profile collection is committed exclusively to use for personal identification purposes, several policy choices present themselves: (1) Should the system be designed to support any type of research involving the stored information? (2) Should the system be designed to aid in the identification of the sources of new DNA samples without clear matches in the database?, and (3) Should the system be designed to support electronic dragnet screening of the population in search of particular individuals? In the context of the expanding uses of iDNAfication, these choices raise some important social policy issues that go well beyond issues of personal privacy.
RESEARCH USES. Among the legislatively authorized uses of the existing iDNAfication databanks is their use for various kinds of research. For example, many state statutes, following the FBI's legislative guidelines, provide for the use of convicted offender iDNAfication data in research by state forensic scientists designed to improve iDNAfication techniques and protocols. Although the state statutes vary widely in the security procedures they mandate for containing this research within the crime laboratories and protecting the identities of the sample sources, if they were to implement the protections recommended by the FBI (Murch and Budowle) using such samples would raise few direct privacy issues. However, it is worth noting that to the extent that this research requires access to physical DNA samples, it provides the main impetus for retaining samples in state crime labs after the database profiles have been generated. This opens the door for other research uses of the collection. For example, Alabama allows the use of anonymous DNA samples from its convicted offender collection "to provide data relative to the causation, detection and prevention of disease or disability" and "to assist in other humanitarian endeavors including but not limited to educational research or medical research or development." (Alabama Laws  1st Spec Sess Act 94–100).
Alabama's generosity towards researchers is presumably premised on the view that the anonymity of the samples provides adequate protection of the sources's privacy, and frees the state from having to worry about the usual elements of biomedical research, like informed consent. But on the contrary, from the perspective of research ethics, these samples are not anonymous, nor even anonymized, since the iDNAfication database is itself the key to identifying the source of any given sample. Since that existing linkage makes it technically possible to benefit and harm the sample donors with the results of such research, all the usual biomedical research protections should apply (Clayton et. al.). In addition to these personal privacy issues, moreover, open-ended research on iDNAfication samples also poses broader questions of research justice. Collections of DNA samples from criminals or soldiers, for example, are likely to be perceived as particularly rich research resources by those interested in studying genetic factors involved in antisocial or aggressive behavior. Unfortunately, our social experience with such research has not been good (Marsh and Katz). Repeatedly, such studies have succumbed to ascertainment biases that ultimately mischaracterize—and stigmatize—groups of people that are disproportionately represented in the systems under study for social reasons. Two forms of injustice tend to flow from these results. First, genetic claims about individual research subjects, like those concerning XYY syndrome in the 1970s, become generalized to an entire class, simultaneously pathologizing behavior and stigmatizing bearers of the genetic trait. This has the effect of both undercutting personal responsibility and legitimizing draconian medical responses to the targeted behavior, like eugenic sterilization. Second, genetic studies tend to misdirect attention from the overwhelming social causes of the behaviors they purport to explain, by encouraging a determinism that suggests that efforts at social reform are ultimately futile. Where this misdirection reinforces existing social policy inequities, it is likely to have an even more pronounced effect (Wasserman).
PROFILING USES. The third kind of databank that is part of a comprehensive iDNAfication system (in addition to the identified DNA profile collection and the aggregate population polymorphism frequencies database) is an open case file: a collection of DNA profiles taken from crime scenes or battlefields or plane crash sites that come from as yet unidentified sources. Obviously, this collection needs to be comparable to the identified reference collection, which means the same markers should be used to compose the profiles in both. With these collections, however, investigators will be especially tempted to glean as much information as they can from their genetic analyses in their efforts to compose a profile of their missing sample source. One of the areas of highest interest has been in non-coding polymorphisms that would allow investigators to estimate the ethnic affiliation of a sample source (Shriver, et al.). These investigators call their markers population specific alleles (PSAs), and the ethnic populations they mark are, once again, just our traditional races: European-Americans, African-Americans, native Americans, and Asian Americans. Should these PSAs be included in or excluded from the panel of markers established for our universal, humanitarian iDNAfication system? Including them would allow the system to support an open case file that could take advantage of the additional information to narrow the search for sample sources. It would also, presumably, take the guesswork out of deciding which racial reference group to assess a particular sample against.
Of course, including PSAs in iDNAfication profiles would elevate the informational content of the profile beyond that of a traditional fingerprint, constituting more of an intrusion on privacy. Moreover, it would do so by reporting a particularly socially sensitive feature of the arrestee: their probable race. But photographs also can reveal race, and we sanction collecting them for identification purposes. How would this be different?
Photography is an illuminating analogy here. Photographs show only the superficial distinctions that we use socially to categorize a person's ethnic affiliation. They leave that categorization itself up to the observer, and make no claims about its merits. Thanks to our large-scale hybridization, in other words, passing for one race or another is still possible in mug shots. PSAs, on the other hand, are defined in terms of our society's racial categories, and purport to be able to appropriately classify even interethnic individuals into their true (ancestral) categories.
This has several implications. First, it means that genuine secrets might be revealed through PSA screening: for example, shifts in the social (racial) status of the arrestee or her ancestors that have nothing to do with their arrest, but which, if interpreted as normative, could cause psychological and social harm to the individuals and their families by upsetting their social identities. In that sense, PSAs are more threatening to privacy than photographs. Second, as the scientists's own hopes for appropriately classifying hybrids shows, it is hard not make the logical mistake of moving from the use of social categories to define the PSAs to then using PSAs to define our social categories. This mistake raises two important issues about the use of PSAs in iDNAfication schemes.
First, it risks exacerbating racism by reinventing in statistical and molecular terms the arbitrary social apparatus of the blood quantum and the One Drop Rule: Under PSA screening, one's proportional racial endowment could be quantified, and carrying the defining polymorphisms for any given race would warrant (statistically) affiliating one with it for official identification purposes, regardless of one's superficial social identity. In the wake of a program of iDNAfication in which thousands of American's would have their PSAs determined, this could have powerful social consequences. In fact, our bad experiences with other forms of low tech racial profiling in law enforcement has already led to court decisions prohibiting the practice as unconstitutional under the Equal Protection clause (Johnson).
The second danger in estimating ethnic affiliation through PSAs is the way it facilitates the reification of (fundamentally unjust) social categories as biological realities. If PSAs are not genes for race, they are at least differentially associated with the people we classify in particular races. Genetic association, however, in the public and scientific mind, often comes to imply causation that implies in turn the objective reality of the effect. In other words, if PSAs correlate with racially defined populations, they must be linked somehow with the defining genes of those populations, and if the racial populations have defining genes, races must be real and separable biological entities, not just social constructions. Our society has had recurrent experience with this kind of hardening of the categories, all of which has been detrimental to the least well off (Duster) because it fosters a particular form of social harm: the erosion of our sense of solidarity as a community and our empathy for members of other groups, leading to what one scholar has called social policies moral abandonment (Wasserman). Any widespread iDNAfication program that involved PSA-based ethnic affiliation estimations would run the real risk of exacerbating that harm, by fostering the public perception that PSA-based profiles revealed real racial assignments.
DRAGNET USES. Finally, there is a third set of choices about the range of use to which any arrestee iDNAfication system should be put. Given our commitment to the presumption of innocence, should such a system accommodate sweep searches of its stored profiles in the pursuit of a criminal suspect? Obviously, in addition to the precise identification of sample sources, the principal purpose of the existing convicted offender iDNAfication databanks in law enforcement is to aid in the identification of suspects by matching unidentified DNA samples from a crime scene with an identified profile in the collection. If in fact we kept the informational content of arrestee iDNAfication under the pattern matching standard of manual fingerprinting, could we really complain about police searches of arrestee iDNAfication databases for the same purpose?
On one hand, it is clear that some dragnet uses of iDNAfication would not be acceptable in the United States. Critics of current forensic iDNAfication programs often point to the 1987 British case in which every male resident in three Leicestershire villages was asked to voluntarily provide a DNA sample to the police in an (ultimately successful) effort to identify a murderer, as an cautionary sign of things to come (Wambaugh). However, given the coercive nature of such a request (police made house calls on those failing to appear for sampling), its effect of shifting the presumption of innocence to one of guilt, its lack of adequate probable cause, and the U.S. Supreme Court's rejection of similar uses of manual fingerprinting, it seems implausible that such a sampling practice would be constitutionally sanctioned in the United States.
However, what if the dragnet were only a matter of searching a database of DNA profiles previously collected by the state for the identification of arrestees? In supporting the existing convicted offender iDNAfication databases, the courts have argued that the public interest in prosecuting crime outweighs any presumption of innocence that criminals may have in future cases, thus justifying the reuse of their DNA fingerprints for forensic matching (Jones v. Murray, 1991). Moreover, we already store and reuse arrest photographs and manual fingerprints, even from those arrestees subsequently cleared of their charges, in attempting to identify suspects in future cases. Why should arrestee DNA fingerprints be handled differently?
Here is where the uniquely biological side of iDNAfication reenters the analysis, with its increased claims of physical privacy. U.S. courts have ruled that systematic analyses of tissue samples and body products (as opposed to photos and fingerprints) of suspects (as opposed to convicted criminals) are the sorts of searches that are protected by the Fourth Amendment, even when the samples are already in the state's hands. This suggests that, although one's arrest presumes enough probable cause to justify sampling for identification purposes, arrestees have not forfeited as much of their presumption of innocence and the physical privacy that attends it as convicted offenders have, whose samples can be searched at will by the state. If these decisions are accepted as precedents for iDNAfication, efforts to screen forensic DNA against a database of arrestee profiles from citizens who have no convictions would also have to pass the Fourth Amendment's tests, and show probable cause for each attempted match.
Moreover, if anything, the bar to dragnet searches of arrestee iDNAfication collections should be set higher than the bar to searching other tissue samples and body products, because DNA profile matching actually poses a greater risk to privacy than other forms of tissue typing. This is because, unlike both fingerprint and urinalysis screening, the process of matching a forensic sample against an iDNAfication database can reveal familial relationships as well as identities. Unlike fingerprints and photographs, in which the environmental vagaries of human development usually work to obscure any convincing evidence of kinship, DNA profiles can demonstrate those relationships in clear genetic terms.
Thus, when non-coding nuclear DNA markers are used to profile a forensic specimen, the siblings, parents, and children of the specimen source will all show partial matches with the specimen. Their appearance in an arrestee iDNAfication database will not make them direct suspects, because of the mismatching elements of the profile. But their matching elements can reveal that they are related to the suspect, and so will flag their family for further investigation by the police. Moreover, when mitochondrial DNA is used for genotyping, the resulting profiles will almost always be completely shared by the DNA source's mother and siblings, and by her mother and all her siblings as well: They are all essentially mitochondrial clones. In these cases, the appearance of family members in an arrestee database might even make them immediate suspects for investigation. In any case, the disclosure of the identities of a suspect's relatives is not something that fingerprint searches accomplish, which means that iDNAfication puts more personal information at risk. It therefore poses a greater threat to the privacy of both the arrestees and their kin. Moreover, experience from clinical DNA testing within families demonstrates that even in a supportive context, the disclosure of familial relationships can have tremendous psychosocial impact on family members (Juengst). To have those relationships disclosed publicly in the context of a criminal investigation only amplifies the risk that the impact will be negative on both the sample sources and their kin.
It is interesting to note in this regard that some states's convicted offender iDNAfication databanking statutes already include provisions mandating the expungement of a person's DNA profile, and the destruction of their samples, if their convictions are overturned or dismissed on appeal (McEwen and Reilly). The only circumstance in which that this happens with traditional fingerprints is in case of juvenile acquittals, where expungement is justified in terms of the burden of an early criminal record on the life prospects of the acquitted. This suggests that having one's DNA on file with the state is also recognized, at least in some states, to carry privacy risks to the individual that are unfair to impose on citizens cleared of criminal guilt, in the same way it is unfair to impose a criminal record on a reformed youth. But if that is true of those whose convictions are overturned, it should be equally true for those who are never convicted in the first place (Nelkin and Andews).
eric t. juengst
Bereano, Philip. 1990. "DNA Identification Systems: Social Policy and Civil Liberties Concerns." International Journal of Bioethics 1: 146–155.
Chakraborty, Rajit, and Kidd, Kenneth. 1991. "The Utility of DNA Typing in Forensic Work." Science 254: 1735.
Clayton, Ellen Wright, et al. 1995. "Informed Consent for Genetic Research on Stored Tissue Samples." Journal of the American Medical Association 274: 1786–1792.
DeGorgey, Andrea. 1990. "The Advent of DNA Databanks: Implications for Information Privacy" American Journal of Law and Medicine 16: 381–398.
Duster, Troy. 1992. "Genetics, Race and Crime: Recurring Seduction to a False Precision." In DNA On Trial: Genetic Identification and Criminal Justice, ed. Paul Billings. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Hoyle, Russ. 1998. "The FBI's National DNA Database." Nature Biotechnology 16(November): 987.
Jeffries, Alex J.; Wilson, V.; and Thein, S. L. 1985. "Individual-Specific Fingerprints of Human DNA." Nature 316: 76.
Johnson, Erica. 1995. "A Menace to Society: The Use of Criminal Profiles and Its Effects on Black Males." Howard Law Journal 38: 629–664.
Jones v. Murray, 763 F. Supp 842 (W.D. Va 1991).
Juengst, Eric. 1999. "Genetic Testing and Moral Dynamics of Family Life." Public Understanding of Science 8(3): 193–207.
McEwen, Jean. 1995. "Forensic DNA Data Banking by State Crime Laboratories." American Journal of Human Genetics 56: 1487–1492.
McEwen, Jean, and Reilly, Philip. 1994. "A Review of State Legislation on DNA Forensic Data Banking." American Journal of Human Genetics 54: 941–958.
Marsh, Frank, and Katz, Janet, eds.1985. Biology, Crime and Ethics: A Study of Biological Explanations for Criminal Behavior. Cincinnati, OH: Anderson Publishing Company.
Murch, Randall, and Budowle, Bruce. 1997. "Are Developoments in Forensic Applications of DNA Technology Consistent with Privacy Protections?" In Genetic Secrets: Protecting Privacy and Confidentiality in the Genetic Era, ed. Mark Rothstein. New York: Oxford University Press.
National Academy of Sciences. 1992. DNA Technology in Forensic Science. Washington, D.C.: National Academy Press.
Nelkin, Dorothy, and Andrews, Lori. 1999. "DNA Identification and Surveillance Creep." Sociology of Illness and Health 21: 689–699.
Rabinow, Paul. 1992. "Galton's Regret: Of Types and Individuals." In DNA On Trial: Genetic Identification and CriminalJustice, ed. Paul Billings. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Sankar, Pamela. 1997. "The Proliferation and Risks of Government DNA Databases." American Journal of Public Health 87(3): 336–337.
Scheck, Barry. 1994. "DNA Data Banking: A Cautionary Tale." American Journal of Human Genetics 54: 931–933.
Schulz, Marjorie. 1992. "Reasons for Doubt: Legal Issues in the Use of DNA Identification Techniques." In DNA on Trial: Genetic Identification and Criminal Justice, ed. Paul Billings. Cold Spring Harbor, NY: Cold Spring Harbor Press.
Shriver, Mark, et al. 1997. "Ethnic Affiliation Estimation by Use of Population Specific DNA Markers." American Journal of Human Genetics 60: 962–963.
U.S. Congress Office of Technology Assessment. 1990. Genetic Witness: Forensic Uses of DNA Tests, OTA-BA-438. Washington, D.C: U.S. Government Printing Office.
Wambaugh, Joseph. 1989. The Blooding. New York: William Morrow.
Wasserman, David. 1995. "Science and Social Harm: Genetic Research into Crime and Violence." Philosophy and Public Policy 15(Winter): 14–19.