Family Tree DNA was criticized for secretly working with the FBI. Now it’s explicitly asking potential customers to help law enforcement.
Give us your DNA. Help catch a criminal. That’s the message of a recent ad from the genetic-testing company Family Tree DNA. The video stars Ed Smart, whose daughter Elizabeth Smart was abducted at age 14, exhorting viewers to upload their DNA profiles to the company’s website.
Not so long ago, DNA-testing companies were known only for their promise to unlock medical secrets or trace family histories. What’s changed is the arrest of the alleged Golden State Killer. Since police tracked down a suspect in the notorious case by uploading crime-scene DNA and finding distant relatives on a genealogy website, the same technique has led to dozens more arrests for rapes and murders. Forensic genealogy has become, if not exactly routine, very much normalized.
The Ed Smart ad is, implicitly, an argument that consumer DNA databases should be used for law enforcement. Family Tree DNA came under fire in January when BuzzFeed News revealed that the company had been quietly working with the FBI. Family Tree DNA sells at-home DNA test kits, but it also allows people to upload genetic profiles from competitors such as 23andMe and AncestryDNA to its website. In late 2018, the company’s CEO later told Forensic Magazine, it discovered that the FBI was trying to upload genetic profiles from crime scenes. Rather than fighting the FBI, Family Tree DNA changed its terms of service to allow law-enforcement use in cases of “violent crimes”—without notifying its customers, until BuzzFeed News started asking.
Genealogists were shocked that Family Tree DNA would keep this secret. (Investigators in the Golden State Killer case and others had used the same methods with another genealogy database, called GEDmatch, which became aware of their involvement at the same time as the public.) But on the underlying question of law enforcement using genealogy databases at all, genealogists have had fewer qualms. A poll of 639 genealogists by Maurice Gleeson last year found that 85 percent were “reasonably comfortable” with law enforcement using GEDmatch to identify serial rapists and killers. And in October, bioethicists at Baylor College of Medicine published the results of a more generalized survey: Of the 1,587 respondents, 91 percent supported forensic genealogy for violent crimes, and 46 percent for nonviolent crimes.
So instead of backing off, Family Tree DNA appears to have leaned into the controversy. (The company did not immediately respond to a request for comment.) While other major players in DNA testing, such as 23andMe, AncestryDNA, and MyHeritage, have resisted law enforcement, Family Tree DNA now allows investigators to upload the suspect’s DNA profiles to find potential relatives. Their access to full DNA profiles of everyone in the database is restricted, though, and customers can opt out of law-enforcement matching. Forensic Magazine reports that less than 1 percent of U.S. customers chose to opt out after one week. GEDmatch did not see an exodus of users after the Golden State Killer case either.
Americans, it seems, are not that concerned about sending a relative to prison. In most cases, the suspect’s DNA profile will match only distant relatives, such as second or third or fourth cousins, who might not even know each other. (A distant-relative match, in combination with social media and public records, can be enough to ultimately ID the suspect.) A woman in Washington State recently found out her DNA on GEDmatch had led to the arrest of her second cousin twice removed for murder in Iowa. Before she shared her DNA, her brother had worried about getting a family member arrested. But now, she told The Gazette of Cedar Rapids, “I feel OK about it … I want someone to have to do time if [he/she] did something like that. I don’t regret it now.” She had never met and did not know the man arrested.
Christi Guerrini, an ethicist at Baylor who co-authored the October survey, told me she had been surprised to see such high support for law-enforcement use of genetic genealogy. The survey is prefaced by a description of the Golden State Killer, and she acknowledged that mentioning such a “notorious” case could bias respondents.
On the other hand, that is exactly how the American public as a whole was introduced to forensic genealogy. Genealogists told me they feared a public backlash to other possible test cases—such as identifying a baby abandoned by his mother. Arresting a suspected serial killer who murdered at least 13 people and raped at least 50 made the technique a much easier sell.
Erin Murphy, a law professor at NYU, says it’s common to see new, potentially controversial forensic techniques tested in cases that will bring out the most public sympathy. She points to the example of law enforcement building their own DNA databases in the 1990s, which have expanded considerably in scope since then. “DNA databases did not start with collecting DNA from people at traffic stops,” she says. “They started with collecting DNA from repeat sexual offenders and people convicted of serious crimes.” Now some local police departments are using “stop and spit” to build largely unregulated DNA databases.
The victims in the cases solved by genetic genealogy are often very young, often female, and often white. There are practical reasons for this: The victims of rapists who left behind their DNA are likely to be women. The people in genealogy databases are disproportionately white. And the forensic genealogy work can run thousands of dollars, so law enforcement is submitting cases deemed the highest priority. This means that law-enforcement use of genealogy is being sold to us with the victims who arouse the most attention in the media and among the public.
That’s true of the victims of the Golden State Killer, who terrorized California suburbs. It’s also true of Elizabeth Smart, whose blond-haired, blue-eyed face was all over the news before and after she was found in 2003. These are the cases, perhaps, easiest for Americans to get behind—even if it means giving up a measure of privacy.