吃瓜大本营

Police departments around the country are beginning to deploy 鈥淩apid DNA鈥� machines, which can take a cheek swab or other genetic sample and automatically generate an identifying DNA profile. These machines cost as little as $30,000 and claim to take just 90 minutes. They handle some of our most sensitive information (our DNA) 鈥� and yet their design and use is largely unregulated. That should be cause for concern for several reasons.

First, Rapid DNA machines are likely to increase the risk of misidentification and wrongful conviction. Even traditional DNA analysis 鈥� which is typically done by trained professionals according to rigorous protocols in accredited laboratories 鈥� has led to and . For Rapid DNA machines 鈥� which police officers operate in precinct houses, without clear protocols, and after at most just a few hours of training 鈥� the risks of such injustices are even higher. Limited pilot studies have already about the machines鈥� accuracy, including their failure to produce usable profiles, contamination of samples due to leaks in the machine, and the generation of at least one faulty profile.

These machines are also being used by police in ways they aren鈥檛 intended for. They were designed to test samples taken from individuals for identification purposes, but local police departments are already deploying them on crime scene evidence, which is often far more complex. Such samples often include DNA that may be damaged or degraded, is present only in low amounts, or mixes many people鈥檚 genetic material. The FBI鈥檚 expert 鈥淪cientific Working Group on DNA Analysis鈥� that only a trained forensic DNA analyst can interpret crime-scene samples, and that Rapid DNA machines should not be used on them. The scientists also point out that 鈥渃rime scene samples are often irreplaceable, and Rapid DNA instruments the entire sample.鈥� The National District Attorney鈥檚 Association a similar position. But local police are not listening.

Second, 鈥渃heap and easy to use鈥� is a perfect recipe for overuse, particularly when it comes to sensitive technologies in the hands of the government. We have seen this dynamic with cell phone location tracking, face recognition, and communications eavesdropping: intrusive information collection that was once subject to 鈥渘atural limits鈥� because it was expensive gets deployed far too broadly when new technology makes it cheap. Rapid DNA machines are likely to have this effect. Our DNA is far more than a replacement for the fingerprint; it is the 鈥渘uclear weapon鈥� of identifying technologies. It can reveal much more 鈥� and more intimate 鈥� information than simply our identity, including our propensity for certain diseases, our family members, and our ancestry. And, as technology develops, DNA may reveal even more.

Third, Rapid DNA machines are likely to encourage the growth of government DNA databases, putting some of our most sensitive information in government hands. It also raises the likelihood of false hits. Even before Rapid DNA machines, we were seeing local police departments asking too many people 鈥� kids 鈥� to provide DNA samples. This cheaper, faster technology 鈥� combined with the fact that many states have expanded their laws to authorize DNA collection from those who have merely been arrested 鈥� is only likely to exacerbate the problem. According to reports by the and outlets, the FBI is working to enlist these machines into a national pipeline that will pour into its centralized DNA database, CODIS. And these machines are also likely to incentivize the growth of , which are maintained at the local level with far fewer quality, privacy, and security controls than federal databases.

Fourth, the easy availability of DNA testing through these machines is likely to exacerbate existing problems with the criminal justice system, including racial disparities in DNA collection that exist because our criminal justice system disproportionately suspects, arrests, and convicts people of color 鈥� and collects DNA from them accordingly. Rapid DNA machines may also encourage police attempts to obtain DNA without a warrant, including through pretextual arrests and the collection of so-called 鈥渁bandoned鈥� DNA that we all scatter around as we live our lives (for example on the things we touch and eat).

Recommendations

Given these risks, policymakers at all levels should think hard about whether it is appropriate to allow police departments to deploy Rapid DNA machines. Where police departments have already acquired them or been permitted to do so, measures (some which we have discussed previously) must be taken to guard against the pitfalls discussed above, including:

• Strict quality controls. The FBI is currently use standards for Rapid DNA machines that will feed into CODIS. But local police departments are already using the machines outside of any standards. State lawmakers should look at this issue promptly and impose quality controls, including training and validation standards, to prevent sloppy use from creating injustices.
• Formal acquisition, use, and retention restrictions. This technology demands formal, legally binding restrictions on the collection, use, and retention of DNA. This should include regulations that: limit use to individuals who police have probable cause to believe committed a crime; ensure that the samples and results will be used only for their intended purpose; guarantee individuals whose DNA is collected and tested a right to access their results; and ensure expungement as soon as a person鈥檚 DNA data is no longer needed for the purpose that justified its collection.
• Limitation on 鈥渧oluntary鈥� collection. Collection of DNA should not be based on the fiction that submission is 鈥渧oluntary鈥� 鈥� that people are free to refuse police officers asking them for a biometric reading. Police officers have significant power and discretion in their encounters with us, and few such encounters are free from coercion. The New York Times story clearly showed such abuse in action; in the Pennsylvania town they profiled, where police must obtain consent from those under arrest, 90 percent of individuals asked agreed to hand over a sample. Officers explained that away with 鈥渃riminals do stupid things,鈥� but more likely most of those people did not feel they had a choice. Even people with more privilege may fear saying 鈥渘o鈥� to a police 鈥渞equest鈥� that sounds an awful lot like a demand. No one should have to identify themselves to the police if they鈥檙e not suspected of a crime 鈥� and certainly should not be asked to turn over their DNA.
• Democratic control. Many local departments purchase surveillance technologies using DHS or DOJ grants that circumvent the local democratic budgetmaking process. And at least one town by ProPublica paid for its DNA database with funds acquired through (aka 鈥渃ivil asset forfeiture鈥�). No police department should spend money on this technology 鈥� no matter its source 鈥� without permission from its city council, county board, or other democratically elected oversight body. That is true even in jurisdictions that have not yet passed the 鈥淐ommunity Control Over Police Surveillance鈥� (CCOPS) legislation that the ACLU recommends.
• Transparency. Democratic control over police technology is impossible if communities don鈥檛 know what their officers are using and how. Legislators should ban nondisclosure agreements with technology vendors in which the government promises not to be transparent to its community about the nature of the technology it鈥檚 buying. Among other things, such agreements interfere with independent expert study of the technology. Individuals accused of a crime should also be guaranteed access to any available information about their tests and how they were conducted.