 for coming here tonight. My name is Tori Bosch and I'm the editor of Future Tense, which is a partnership of New America, Arizona State University and Slate Magazine. We look at emerging technologies and their implications for public policy and for society. We do that both through live events like this and through a channel on Slate at Slate.com slash Future Tense, where we also publish three pieces related to tonight's conversation. I've been editing Future Tense for close to eight years and the thing that constantly strikes me is that the future is so much farther away than it appears. So many of the conversations that we were having when I first started this job in 2011 are ones we're still having. Even something like self-driving cars. We've had some changes, but it's not like you can now just go out and buy a self-driving car. So many of the times these things move much more incrementally than we expect. But a huge exception to that is what we're going to be talking about tonight. Law enforcement use of open genetic databases filled with information uploaded by people who took consumer DNA tests and using it to solve crimes. It's a genuinely new development, something that the public first learned about last year with the announcement that they had solved the case of the Golden State Killer. Now, your DNA test could convict a third cousin you've never met or your third cousin's DNA test could convict you for a crime. Tonight, we're here to discuss the ramifications of that and how to balance genetic privacy with a genuinely useful tool to solve crimes that people thought could never be solved. Our speakers will speak for about 45 minutes and then we'll open it up for questions from the audience. If you have a question, please be sure to wait for the microphone because we're also streaming the event live tonight. Afterward, we invite you to stick around, have another drink and continue the conversation amongst yourselves. So now I'd like to introduce our moderator who will also introduce our speakers. Rebecca LaVoy is New Hampshire Public Radio's digital director overseeing NHPR's website, online news, and on-demand content, including the podcast, Barebrook. How many of you have listened to Barebrook? Yes, it's the best. It was my favorite podcast of 2018. She is also the host of the hit podcast, Crime Writers On, as well as co-host of the completely unrelated slate podcast, Mom and Dad Are Fighting. So Rebecca, I'll let you take it from here. Hi. I am really excited to be here. One thing that was not on my resume just now is I'm also a true crime author. I have experience reporting and writing about cold cases and the extensive shoe leather work that goes into solving them. Sometimes people come forward after 15 years just wanting to tell a different story than the one that they told. And there's just a lot that goes into criminal investigations and the intersection between that and rights and civil rights and criminal justice is an area of particular interest to me, which is why when this technology first sort of hit my radar, which was before the Golden State Killer happened because I happened to work with Jason and he's been reporting on this for two years, it really struck a chord with me. So we have an amazing panel and I'll introduce them. John Fitzgerald, he is the Chief of Police of the Chevy Chase Police Department and I've also discovered a former attorney. I like it. But I'm still in therapy. Yes. I'm getting better. He also represents the, he's on behalf of the Maryland Chief of Police Association and the Maryland Sheriff's Association and we'll talk about his advocacy there in a second. Also we have Nila Bala, she is the Associate Director of Criminal Justice and Civil Liberties and a Senior Fellow at the R Street Institute and she has also worked as a public defender in Baltimore. And we have Jason Moon, the reporter and host behind Bear Brook, a case, a criminal, a criminally, a critically acclaimed podcast about a decades long old cold case that actually gave birth to the technology we're gonna talk about. So I think I'm gonna start with Jason. Genetic genealogy. Can you, as you did on the podcast, for anybody here who might not understand the difference, say, between the DNA we see that I'm using in law and order to catch people and this technology, can you give a thumbnail sketch of what it is, how it works? Yeah, so if you think about what genealogy is, that's the construction of family trees, relationships between people, lineages and that traditionally has been based on records. So birth and death records, sobituraries, wedding announcements, et cetera. Genetic genealogy is that same system sort of supercharged with some really sensitive new DNA tests that are available now. And so the way this is different than what you might see on a law and order episode where you have a crime scene sample of DNA and you have a suspect, that test is to see if those two things are the same if they match. A genetic genealogy test, you could have a crime scene sample and you're not looking for a match necessarily but say a half a percent match in a genetic genealogy database because that means you found that person's fifth cousin and then from that fifth cousin you can use the familial connections on the family tree to figure out who the owner of that sample is. So it's a sort of exponential advancement in the usefulness of DNA evidence for crime scenes to find suspects but also to identify missing people and I think we're trying to figure out how many cases have been solved with this by now and it's losing count, it's in the dozens. That's right. And so can you just briefly explain the players? You have, it's not 23 and Me and Ancestry who are these open databases and it's not a police department who is running this test, right? Yeah, that's probably the most important caveat to point out is that if you bought your grandma a kit on Ancestry.com her DNA is not being, her police are not using her DNA to solve crimes. So until recently it was just one database called GEDmatch which is an open source, sort of it's not a private testing company in that you don't send GEDmatch your tubes of spit. One of the main reasons GEDmatch became popular was for people who tested on different sites and who wanted to compare. So if I tested on Ancestry and Rebecca tested on 23 and Me before GEDmatch if we wanted to compare our results one of us would have to buy a new kit and wait for the tube and spit in it and pay 100 bucks or whatever. GEDmatch allows you to sort of export the digital version of your DNA results and then compare them directly for free. So it became pretty popular about a million and some people went on it but the rules for who could access it were different partially because it was a digital database and not a physical collection of tubes of spit which is what 23 and Me actually has or Ancestry has. Now there's two databases that allow police to use it, the other one at Lewis Family Tree DNA. And so those are the two, if you're thinking about law enforcement using them to solve crimes or to identify missing people, it's those two databases, not the really big ones. Though some people would love for those really big databases, Ancestry 23 and Me to open the door so long enforcement. Other people have serious concerns about that which I'm sure we'll get into. That's why we're here. So this is opt-in. You have to decide to send your profile. However. It is, I would put an asterisk on that though, just to, and skeptics would say, well you didn't know if you put your DNA on GEDmatch in 2015, it'd be pretty hard for you to have known that that was a possibility, that your DNA might a couple years from now be used to solve a criminal matter. Since this stuff has happened, GEDmatch has updated their terms of service and now it does explicitly say, when you sign up, that it could be used for this. But there's, some people would argue that terms and conditions, forms aren't really read anyways and maybe they're not the most efficient way to get permission for these sorts of things. Right, because we all read them thoroughly when we click on our iTunes. So the rub here, and this is the thing that has a lot of people talking, is that while you may opt in, do your 23andMe test and then upload your results to GEDmatch, you're not just opting yourself in. You're opting everybody who shares any of your DNA in as well. Not just siblings, but second cousins, third cousins, distant relatives, future children. Nilaad, can you just talk a little bit about what you do? And I'd love to just give you a little background on when you're coming at this, where you're coming at it from, given the work that you do. So like Rebecca said, I'm now sort of in the policy world looking at criminal justice and civil liberties issues broadly, but my prior career was as a public defender, both in Baltimore and in Santa Clara County, California. So when I look at this issue, I really think about the court system and the players in the court system and the privacy concerns. And to me, I think it's important to remember that when we put privacy and public safety head-to-head, and in a case like the Golden State Killer, it's easy to go for the public safety. You have a serial killer, a rapist, you want them to be brought to justice. So I'm really sympathetic to that, so I want that to sort of be there at the outset. But at the same time, it's really important to remember that privacy is necessary for a civic society. It's part of what makes us free, the fact that we can control the information that the state knows about us, the fact that we have private spaces where we can form our own identities or personalities, develop as fully autonomous human beings. So I think Rebecca just hit on the point that makes this issue so hard, that normally we do control our own privacy, we can decide when we want to share information and if we share it with third parties, the Supreme Court essentially said, well, we don't have an expectation of privacy, we've decided and consented to that. This is different for two main reasons. One is we're actually consenting perhaps on behalf of a lot of other people, and the other is kind of what Jason was mentioning, maybe when we consented we didn't expect the information to be used in this way. A lot has already been written about the privacy of the person who uploads from 23andMe to this open source database. So I think there are things that we can think about, for example, making sure that the disclaimers are stronger or Family Tree DNA actually just came out this week and said they'd create an opt-out feature. So you could actually click a box and say, you don't want this information to be used by law enforcement. So those are all I think progressive things, perhaps in the right direction, to give people more control over their own information. But as someone who's worked in the court system and kind of seen how criminal cases progress, albeit not any cases I think where this genetic genealogy was used, I'm just skeptical and sort of wary of just the release of this technology without any process. So I think we'll find that even though some of my views might be counter to what John will say, at the end of the day, it's not that I don't want this technology to be harnessed when appropriate, it's just that I want there to be process. And we can talk about what that process might look like, whether it's a court order or a warrant or something. And the reason I think that's so important is because there are huge privacy concerns that extend beyond the length of a given case. Three things I'll quickly highlight is first, there can be mistakes made, right? Human error does happen. And I think when we think about DNA technology, sometimes we think it's infallible because it's technology, it's science. But there's human intervention at every stage of the process. And so I'm thinking about perhaps an instance where DNA is collected at a scene and uploaded onto this open source database, but it ends up not being the DNA of the suspect in reality, it's somebody else's. And now that person's information is exposed and they don't even know that their information has been uploaded to this database. So unlike the individual who actively participates in 23andMe and uploads it, this person's position is potentially a little bit different. The second issue I'll highlight is where they do get it right. It is the suspect's DNA, but it's not the kind of Golden State killer case where the public safety issues really outweigh the privacy issues because we have to consider the fact that 80% of state dockets are actually misdemeanor cases. And it might sound far-fetched right now because this technology is still fairly expensive and fairly new, all things considered. But I don't think we're that far from a world where we use this technology more and more. And the last and sort of related issue is that once this is uploaded, I think even if a person like the user, in this case law enforcement who's uploaded it, tries to delete it, it could be living in perpetuity, right? Like when we put things on the internet, on Facebook, even if you delete your account, you can't prevent the fact that somebody uploaded a picture that you uploaded and it's now on their Facebook or on their personal computer or on the internet somewhere. This is a huge issue more broadly with expungements and record-ceiling. Even when we expunge a state database, for example, of someone's crime, that doesn't prevent the fact that there was a newspaper article written about them or a mugshot on a private website that still exists. So I think about the re-entry piece too. So when we do upload someone's DNA, yes, we're harming their privacy and we might be harming it forever. The last thing I'll throw out there is the family members in both of these cases, right? So you have the family members, the people who use 23andMe, they know they're uploading it to Jedmatch, but their family members don't know. And their children, their children who exist now or may exist in the future don't know and the insurance companies that may someday be interested in ensuring their children or not ensuring their children, that DNA is something they might wanna look at. Some of these, again, might sound a little bit in the future, but even though Tori said sometimes the future seems far away, sometimes it really rushes up on us and this is one of those cases. So I'll stop with that. I think, I mean, I can talk about this forever. I think all of us can, because it's so interesting, but. I think our prep call for this panel was the most, we could have just done the prep call as this conversation because we're all that in it. So John, talk about where you're approaching this from and I'd love to hear about, you spoke to me about a bill in Maryland that has you concerned, but I'd love to also get some of your background and you've seen this technology and action solving cases. Right. So for the audience members that may not be quick on picking up clues, I'm a police officer and for those that are listening in audio form, I'm wearing a uniform. But so, this is my 36th year in this work and the only reason and great, this is a great conversation, really good conversation to have. Now, I do differ on some of the things that Nilla said, of course, but it's really important to have this conversation. While Nilla said that she threw some sugar in my direction about wanting to catch bad people that do bad things, I'll send some back her way and that is that I'm also very interested in privacy and keeping government from overreaching. I'm into that. I mean, I have sons and I'm a citizen myself and I don't want to be part of an overreach and nor do I want to be overreached into by government or anybody else. So I'm with you on that or if that's where you are. So I'll give you a little context. Perhaps the only reason I'm here and right now I'm the chief of police of a very small little municipal police department. I don't have a big kingdom but I have a lot of experience. I was in a large organization for a full career and I'm in a second career. So a couple of things that might be helpful for you to understand how this works by dovetailing a little bit on what Nilla and Jason said. So going to one of the things that Nilla said a minute ago about let's, I'm gonna use some terms, evidentiary DNA or forensic DNA. That's the DNA left by a bad person on a crime scene. So evidentiary DNA is obtained by the investigators that are in this case will say it's typically a really serious felony like a sexual assault of some sort or a homicide. So that DNA, and I think most reasonable people can agree that the bad person doesn't have an expectation of privacy in that DNA that's collected as a result of him committing this crime and leaving the scene. So that DNA, if it's then subjected to this laboratory analysis and then there's a genotype file created, it's digital, then that could be uploaded to a public database like GEDMATCH or now Family Tree. So I'm gonna talk about the privacy of the bad guy and then we'll talk about privacy perhaps of the volunteers or family members. They're not the same. So the privacy of the bad guy, when that DNA file is loaded, it is loaded in private. So what do I mean by that? When it's loaded, the person who volunteers can set the settings. It is set as a one way looking glass so that the investigators or in this case the laboratory and while it is paraben nanolabs, it does the predominant amount of work here, they are acting as the agent of law enforcement. I don't wanna absolve law enforcement for having their hands in this. When we hire somebody to do something on behalf of the state, it's the state doing it. So whatever rules might apply to the police would arguably apply to a private agency acting on behalf of the police. So when anybody loads their profile, they can say, I wanna see as much as I can see but I don't want anybody to see me. That is the way this goes. So the person who, the bad guy, even if it is the right bad guy is not exposed. The bad guy's DNA is looking and pinging off of everyone who has allowed their DNA to be seen. That's a decision that has to be made individually. There are perhaps our members that have uploaded their data in the same way and the bad guy's DNA can't ping off of it. It can't see theirs because they've elected privacy as well. So we protect that even though they're not a particularly sympathetic person here, we still protect it and we're going in and just looking outward and not allowing people to look in. That's one thing. As to what do we learn? And this is gonna sort of lead us into one of the other points that Nila made about privacy for family members or people that did not actually upload DNA. So when, let's suppose there was a match. What do we get? How does that look? What does that, how does that work? So let's suppose there are people that have decided deliberately to take their files from 23andMe or from Ancestry and then send it electronically to Jedmatch and they have decided to be public because they might wanna know. The reason for that might be that, hey, I wanna know if there are relatives out there that I haven't met yet. Sometimes you learn that you're the parent of someone you never met, right? It's not always what you expect. So the fact that you didn't expect to find a new child of yours wasn't your intent but sometimes that happens, right? But the point is it's essentially recreational. I wanna find out who I came from and where I came from. Well, if there is a match, it could be many matches. There could be a whole family group that decided to do this together. What comes back to the person who signed up, whether it's you or me or Parabon Labs, they get whatever the person put in to identify themselves. They could put their name and their address on their phone number, they could put an alias, they could put an email address and that's it because they don't know what creeps are out there, who knows, right? So it is only that. We don't see the DNA file of anybody that's uploaded. We see you have a match or you have three matches and they are ranked. The closest relation is this person, ABC at gmail.com. That's all we know. Then we have to work and find out who that person may be. So the work of this science is absolutely DNA-based but the privacy of the suspect, if you will, is not exposed and the DNA of the volunteers is likewise not exposed. Two letter detective work to put that together. I mean, this is something that Jason explored in Bearbrook that, I mean, I'm just going to just touch up right there because like, Jason, can you just talk about, I do wanna come back to this legislation that I think is an interesting topic and sort of trying to strike that balance, maybe not quite the right way. The first time this was used was related to the Bearbrook case. Can you talk about how that happened and then what it took to actually get a result at that point versus now and kind of where we are and how quickly it's evolved? Yeah. It's sort of hard to pin down when exactly this, because there are multiple thresholds that were crossed in terms of the use of this technique sort of creeping into the criminal justice world but as far as I can tell, there's two instances that we cover in the podcast that were sort of the first two thresholds that were crossed and the first one was when they used genetic genealogy to identify Lisa and for people who listen to podcasts, you know, it's super complicated and I'm super simplified that she was essentially a living Jane Doe. She had been abandoned by a man as a child who she thought was her father. She learned in her 20s that he wasn't her father so she had essentially been kidnapped and she had no idea what her real name was or who her family was and so she became a willing participant in a search that used genetic genealogy. It was in fact her idea. This is a victim's idea, told a cop. Why don't you upload my DNA to like Ancestry and see if you can find anybody in my family? I think that's super fascinating. So the detective in question sort of started to look into it then realized it was more complicated than he could handle and so he turned to a community of genetic genealogists who sometimes go by the moniker search angels and who interestingly had their own version of this debate probably 10 years ago, if not more because what they had been using this technique for was to do adoption searches where if I had been adopted as a baby and now I'm in my 20s and I want to know who my biological parents are you could hire or a search angel might volunteer to help you find your bioparents by using your DNA and this even that spurred a debate within the genetic genealogy community about is that appropriate? Is that ethical to be using this to find people who maybe didn't want to be found? If you put your baby up for adoption you can choose in different circumstances to do it anonymously or not and so is this a breach of that trust? I think by and large the side who said we can and should do this one and it became a pretty accepted practice and so that adoption searches is basically the technique that all the criminal stuff was built on top of and so Lisa was the first time that she was basically an adoption search and that they were looking for her bioparents the added step was that it was part of the criminal case because she had been kidnapped, she had been abducted so there was a criminal element to it so that was sort of the first threshold into that world. The second one was in the same case in the Bearbrook cases when they identified the serial killer Terry Peter Rasmussen as the likely murderer behind the four people found in Barrow's in Bearbrook State Park and that was a significant step because he'd identified a criminal suspect though he was deceased. So it didn't lead to an arrest but for the first time we had used genetic genealogy not to identify someone who had been put up for adoption but to identify someone who had committed a crime so that was a pretty important threshold. Then just months later and by the way this is the same genetic genealogist doing all three of these things, Barbara Ray-Venter, just months later she did the same thing to identify the Golden State Killer and that was sort of the third threshold where it was identifying a suspect that led to an arrest where he was still alive and then loose and so I think at that point the seal was sort of totally broken and we are where we are today but that was the sort of evolution from adoption searches to a missing persons case to identifying a deceased suspect to identifying a live suspect and putting handcuffs on them. And I remember when the Golden State, I remember reading, I'll Be Gone in the Dark, the Michelle McNamara book in which she was investigating the Golden State Killer case. I remember being on vacation reading that book. I had just, we were talking about the podcast and it was very much in development at the time. I read the book and I thought this is how this case is gonna be solved and it was just a couple of months later that it was like, it was unbelievable how quickly the technology that I knew about because you were reporting on it that nobody else knew about came to bear in this case. Like all this community is very small, these investigators doing this. Nila, you talk about your sort of fourth party consent and the potential rub there of like the privacy issues of my data's out there and I'm compromising potentially somebody that I don't know and the, but there's also just the basic suspect civil rights part of this that you wrote about in a really interesting piece that's on Slate right now. I think about it too. I think about the carpenter case, the Supreme Court case. I think about many court cases where someone has fished a cup out of somebody's trash and used it to identify them through a DNA match through traditional means. I think about even like, what is it? The Kylo case where they determined that you can't just fly over somebody's house and use infrared to determine who's growing pot and who's not. Because that's in their home. I mean, do you ever want to ask this to question like the thing that your code, your personal code? Do some people believe like that is their, like it's in their home. Like if they haven't been the one to put it up there, like it is still sort of in their own realm to just be out there publicly without them having put a cup in their trash or, you know what I mean? Is that- I absolutely know what you mean, yeah. And I think it's really interesting the cases in sort of the support the amendment jurisprudence. The Supreme Court, I mean, has been trying to figure this out and as technology keeps leaping forward, the jurisprudence that kind of worked to explain the way we experienced privacy may not serve us all that well. I think that's what we're seeing with Carpenter and the cases that will probably come in the following years to give folks some context. Essentially, you know, up until very recently, it was clear that if you were in your home, that was again sort of a sacred space, a private space. But essentially once you left your home, your movements around in the public roads, anything you did on the public streets, that you weren't really considered to necessarily have a reasonable expectation of privacy. The ideas other people can see you, you're exposing yourself, whatever you do. So if the state happens to see you doing something, then well, it's fair game. Carpenter sort of changed that and that was a cell phone case that the Supreme Court considered last year because by itself, the way cell phones work, right, is they ping off towers and they have locational data based on that. And by itself, a person's location, while they're in some public space, it would be okay under everything we knew about Supreme Court, Fourth Amendment jurisprudence before that. But the Supreme Court essentially said, well, we're kind of in a new sort of territory here because when you aggregate all that information together, it paints a very different picture. You can really know so much more about a person's life than just where they were at this one moment in time. And I think this DNA technology is similar in that, yes, a person who is uploading anything from 23andMe to Jedmatch, they are consenting in a sense. They're sharing information about themselves, but the scope of what they're sharing and I think the expectation of what's gonna be happening after that is different from what it's now being used for with criminal investigations. So I think it's a similar issue where DNA is such a private thing, it is such an intimate thing, it's your code, right? So I think the analogy to it being your home is a really interesting one and it will certainly be fascinating to see how I think the courts deal with this. I think the issues that John mentioned are really interesting too and it's not that, that I'm opposed to the way that you kind of described it happening, but my fear is that it doesn't necessarily happen that way all the time right now. Because there isn't a rule, there isn't a rule that law enforcement has to check the one-way mirror box, right? And I think the civically-minded, privacy-interested law enforcement officer perhaps does, but that's what I'm interested in kind of seeing shape out, the laws and the procedures in place so that we know as a society that when they're trying to catch the bad guy, how do they go about setting those parameters? How do they work with these open-source databases to make sure people are protected? How do they make sure they delete accounts after the fact so that people can re-enter society? And when they get it wrong, what's the recourse, right? So those are the things that I think we still need to shape out and whether it's through courts or through statutes, we'll see kind of how this all shakes out. As we talked about before, this is very much on a law enforcement side and I think it's very expensive. So I think about this in the possibilities and I think this could be great for potential exonerees too to have access to this kind of technology to say it opens more windows to finding alternative theories of crimes that didn't potentially exist before, but you also have some concerns around policy. You're not opposed to policy, right, John, but you are opposed to policy that would make it impossible to solve horrific crimes with this technology. Of course, and I think most of it should be. So one of the reasons, and you talked about the bill, I'll circle back to that for a minute because I know you wanna talk about that briefly, there was a bill in this session of Maryland's General Assembly that would prohibit any person, not just the police, but any person from ever comparing and searching a DNA database of any type for the purpose of finding a relative of a suspect. So here's the analog to that. Cars are really, really useful in society. They move us all around. In 2017, 37,000 people died on American roads from cars. Cars also result in harm and death. So what do we do about that? Do we throw cars away? We legislate for driving behavior and for equipment in cars and try to reduce all of that. So while there is certainly an opportunity here for people to do bad things, I think it's important to recognize what is the public good from this? And I would suggest that for anybody that really is thinking, ooh, this DNA is really private. I'll make just a couple of quick points. For those of us that have DNA, can you raise your hand? That would be like all of us. So do you not know that it's kind of personal? Do you not know what it gives everybody to see if you decide to make it public? So these GED matches like the Facebook of DNA data. If you choose, okay, Billy Bad Guy takes a selfie of him and Valerie Victim in the state park. Sends the picture to his aunt and says, Auntie, look, I just met my, I have a new girlfriend. Isn't she beautiful? Sends the picture to his aunt. She uploads it on her Facebook page. Billy Bad Guy didn't know that, didn't give permission. Auntie makes it public. Right around that time, right around that place, she goes missing and a hiker finds her dead body and sure enough, his DNA is at the crime scene. Now, okay. But he didn't expect us to look at that picture. It gave us a lead to figuring this out. And I would say, the other point I would give to you, so that's the expectation of privacy. What is the expectation? How reasonable is it? If you make a choice, but I would agree and with Nilla and I think as a lawyer, it's important to recognize that it's about the knowing, intelligent and voluntary consent. R is a person making a decision that they understand. And if they understand it well enough and if the terms of service or the banner is there loud and clear, then they've decided to do this nonetheless. And I think the expectation goes, the argument of expectation is diminished. I'll say one final thing here on that point. If we talk about the expectation of privacy being that we, the police are kind of, in some way or law enforcement or their agent is digging into data and finding out a lot of things. I can tell you that when most of us, better way to say this, your comfort or your demand for privacy probably enjoys an inverse relationship with how much you love the victim. If you were to talk to a victim that is distant from you, maybe you're more sort of objective and intellectual about it. But if your spouse, your son, your daughter, your mother, your father is a victim of a heinous crime, I would suggest that you would more want science to help identify the person. And if we ultimately do identify the person using DNA science, we do so to the exclusion of all others, conceding the fact that maybe it's not the DNA of the bad guy. But if it is, and we identify that person and hold them to account, not only have we identified the person that we need to identify so that we don't put some of the innocent on death row or in prison, because nobody should ever want that. We identify the right person. We also stop the person from doing this again. And again, as some do. A lot of ifs, though, right? We were talking backstage about the, you are a fair-minded, you listen to podcasts like In the Dark. You are outraged by police abusing their authority. You have a real sense of justice in how it should work. And you also acknowledge there's a lot. You just had like 15 ifs in your... Ah, yes, because the suggestion is okay, there could be bad things, like the car analogy. If the suggestion is get rid of it, I think that we're ignoring something that is extraordinarily important and valuable. And we're only marching ahead in one direction, typically when it comes to technologies like this, recognizing that we need to figure out where the guardrails should go, perhaps. And we need to have these conversations. And hopefully we can work through that. And some of it may be based on legislative initiatives or what have you. But guardrails being important, so is finding some of these answers. So I'm not an apologist for police misconduct or for governmental misconduct. And I, quite frankly, hate it and I've railed against it for many years. But we know that bad things can happen, but we need to focus on, does that mean we throw it away? Jason, you have seen this technology in action, both with Lisa and then with identification of Terry Peter Rasmussen. If you are subscribed to Bear Brook, stay subscribed. There's a reason why I'm saying that right now, trust me. You know what this can do, you've seen it. Where are you? You don't have to give you like your reporter, objective, non-objective answer. I'm just curious, where are you on this after having seen it work? And then with all of the critical thinking that you have as somebody who's been reporting on this and really tearing it apart, probably from longer than any other journalist telling this story. First of all, you did upload your DNA. But you did upload to GenMatch, right? You did the genetic test. I haven't done that. But not for any principled reason, just because I forgot about it. He did it for the podcast, but I wasn't gonna have him listen. I had podcast reasons. I hear him spitting in the cup, it's really. For audio reasons, I did that. I think I'm probably in the middle, to be frank. I think it's undeniable how powerful this technology is. I mean, if you just look at the Bear Brook case and there was a murderer who was unidentified for three plus decades, every conceivable technique, forensic and otherwise had been tried, even some other pretty far out there scientific techniques like stabilisotope analysis, and none of it worked. And then this did in pretty short order. And you look at the Golden State killer case, and again, it's hard to argue the power of these tools. On the other hand, I do think that there is cause for concern. And one way to put this in context, I think, is if you, and I was surprised to learn, or it helped me in learning about this, was if you go back and look at the way that the first DNA tests became a part of the criminal justice system and the investigative system. It was a really slow and measured and sort of committed out process where the FBI was testing the first prototypes of DNA testing on FBI agents for months and years ahead of time. There was like a whole group of 100 FBI agents who were all, their blood samples were like the initial samples that they were testing. And it was a funny story that they did the first test, only 87 of them matched, which meant they just like hadn't figured out how to do DNA testing yet. But the joke was that the FBI Academy was so good that it had changed the DNA of those agents. But then it was slowly introduced and then there were some key cases early on in the late 80s that sort of codified, okay, here's a process, here's what will rise to the level of evidence in a courtroom. And it slowly became this sort of gold standard that we know of it today as where if you do a simple matching technique, it's almost 100% certain. And the way the databases were set up, the CODIS database, the FBI maintains, it's highly regimented in how it can be used, the markers they set up to match, it used to be 13 out to 20 markers. They were specifically chosen so that they could only do a matching test so that they couldn't be used for profiling, that they didn't have any health information about the people, it was a whole thing which is very well thought out because it happens slowly you come in at the time. It was built. Exactly. If you put that next to what's happening today where law enforcement hasn't created this system, they sort of come into a preexisting system and they're using the techniques and the tools and the databases that were built by a subculture of hobbyists basically and have found out that it's really useful for them as well. I think that's where you can see where unintended consequences might come up, where things are happening more quickly than they were the first time around where law enforcement doesn't have nearly as much control over the entire setup of the system and where they're not the only ones who can do this. I mean, that's another thing to remember. It's not just... It's open source. Right, a sort of real investigator could build an API that could scrape the data and figure out patterns of... I mean, there are other potential things that could come from the same data. Right, or a foreign intelligence service might use it to identify Americans abroad. I mean, there's all sorts of... Kind of... You can spend... You can have a lot of fun kind of thinking of like... Minority report stuff. Exactly. It could be a sci-fi novel based on any number of these things. So that's where I think that it's reasonable to sort of have concern. But again, it's concerned because you see how powerful it is. Yeah. Nila, if someone gave you a 23 in May for Christmas or holidays, would you do the test? No, and also, I would probably just find out I'm like 98% South Asia for India. Like, I don't know that I've been at 60% in the end or so as we just got rid of the podcast. Yeah, so I probably wouldn't do it. What about you, John? Yeah, I use a fork. So I might be 50% in the end-earth all, but... So yeah, I don't know if I'd do it. I don't know if I would do it. It's funny, I'm not even on Facebook, right? So I mean, my whole mantra right now is like, I don't want to be out there. I have, as a career law enforcement officer, you don't make everybody happy. I try to be a nice guy to people, but sometimes what you do requires you to get in people's lives. And so, with a family, I kind of keep a very low profile, and I don't know if that would change after the career is over. I don't know. Well, I think we could probably take some questions. What we want to do is make sure everybody goes to the mic because we are creating a video of this presentation. We want to make sure everybody can be heard. So, does anybody have a question? Well, she's got a pad. Robbie has got a question, too. I had her already down, so I wouldn't forget. A little bit for everyone, so whoever feels like they can take on the question, but it's regarding Maryland House Bill 30. And so I understood that bill to mean that it was third-party collection of DNA to convict the accused. So in my example of that would be, police would secretly collect DNA from me to convict my mother. And so we know now that laws have built-in protections against compulsory testimony for special relationships like husbands and wives. So my question is sort of building upon all of that. Do you think, presently, without the type of law that was introduced in Maryland, these special relationships might be compromised? Well, I should speak to that. That's a misunderstanding of House Bill 30. House Bill 30, right now, if it were to pass, by the way, it is not moving forward in committee. It's going to summer session and it won't go away. It'll come back next session, perhaps. If it passed, what it would have done would be it doesn't talk about any collection, whatever. The only collection that is done right now, mandatorily, is done by the state of Maryland for certain felony crimes when someone's either been charged with specific crimes or convicted of it. And that's a mandatory database using about 20 alleles, as Jason mentioned. This is talking about who can look and how can you look? And it would prove, right now, the law as it stands today says that you can use the DNA database owned by the state to compare against a specific unknown sample to see if it is him, if that person, like a fingerprint. If you have DNA from a crime scene, you can put it into the database and say, does the database know him? Is this somebody that's been arrested or convicted before? It can be used for direct comparison and it's quite good for that. It's not at all good for distant matches. And that bill would say, not just that you can't match against the state database for families, that's what the law already says, and I'm okay with that. It's not even a good tool for that. What it says is nobody can use any database for any such query to try to find a lead for someone else who may be out there. And that is the problem with the bill because it is the throw away the entire science regardless of even if the person volunteers. If somebody comes into the police station tonight and says, I saw the news last night and my 17 year old son was the one that was captured on the security camera, he's the murderer. Doesn't always happen, but it does sometimes. Somebody gives up their own family member, they came in, all right? Are we not supposed to listen to that? So if somebody is truly a volunteer, if we really design this or greet guardrails, so everybody's okay with it. We only look to those databases where people are okay with it. It would even throw that away. And that is the problem with the bill. Rabia has a question. Thank you for all your comments and really explaining the issues to us. Does this make okay? Yeah. So my question is for Nilla. And I'm an attorney, I have a healthy skepticism of law enforcement and I have a background in civil rights as well. But to me, I mean, the way it's being described how law enforcement is using it right now or using these databases right now, I'm trying to find like wouldn't a real world example of this be like, for example, let's say a suspect leaves a cell phone at a crime scene. Now that cell phone is registered to an address, maybe not even his own address, but it's registered to an address, to a parent's address or cousin's address, whatever. And it has a list of numbers in it. I mean, so for me, this seems kind of like the same thing. So the police now have all these leads, right? They can go to that address, they can ask people who live there, do you know the person knows his phone? They can call all the people. It kind of seems like the same thing because all they're getting, it seems like, is the amount of information that the voluntary DNA donor agreed to give. If it's just an email address and they can never actually see the DNA profile itself, all they can get is like some basic information and then reach out like they wouldn't, any other law enforcement scenario and say, you might be related to somebody who's a, and they can't compel that person then to cooperate like in any other investigation. So I'm wondering, doesn't that kind of seem like the same type of usage as we see in other investigations anyways with other kinds of evidence? Sure, I understand what you're saying. I think it's different in a couple of ways. One with the cell phone example, I mean, I think that most people do keep their cell phones locked or they're somehow secured. And if they are, then there would be a process to go through in order to access the kind of information that you're saying. And that's all I'm saying. I'm saying let's think of process like we do with most other pieces of evidence. Not that I'm against the police being able to use this information at all, but at the point where they're going to go through the process of actually uploading this person's information into GEDMATCH, just having a third party, like a judge or someone be able to say, what are the privacy concerns? What is the offense here? Like how egregious is the crime and doing a weighing of factors and making sure that all the things that John was mentioning, making sure it is a one-way mirror, making sure there's a provision that this account will be deleted, right? With the cell phone or with almost any other piece of evidence, it doesn't necessarily live on in perpetuity, but a user account on an open source database could, right? So I'm just thinking about all the pieces of the puzzle. And right now I just don't think we've figured out all the things, all the guardrails that need to be in place when law enforcement uses this technology. Whereas with other pieces of evidence, I think we've thought that through a little bit more. May I jump in on that? Yeah, thanks. Again, just really good brain food here. But to go directly at your comment was a really good summation. What I might even say is that most of the time, we're not even reaching out to that person at all. So somebody is a volunteer at a Jed Match and we say, okay, there's a, we've figured out who that person is. We don't want to tip this person off. Especially if they're a close relative. We don't even connect with them. We do standard, good old genealogy work just like it's been done for hundreds of years, but now we have computers and a lot of public records make it much easier. So we start with a name, an identification, somebody we know, and we build the family tree up to the common ancestor, and then back down to the Mr. X. That is the unknown DNA sample. And so with that, we're not even intruding on our life by even a phone call. We don't even want to do that. We want to find our way through this and only identify who this person is. Right, so the great aunt doesn't need to say, yes, you can use me to help catch my unknown nephew. It's just, that's a piece of information that's being used all the time. And I'll be honest with you, I have trouble finding a victim here, and I find victims a lot. I'm having trouble finding a victim in this scenario. Who are we protecting? We're not looking to protect a bad guy here. We're looking to identify him. That DNA is not protected by any constitutional expectation of privacy. The people that, if they knowingly, intelligently, loaded their DNA and said, yes, make it public. And that's the only way we see it. They're not being victimized either. So what they've done is we've looked at only, we've gotten their name as a result of just the same way you would get their name if you were not law enforcement and you uploaded your DNA and you were related to them. Same exact type of thing. And with that information, it's a lead generator only. We work it up and then work it back down. In the case of Lisa, I think it was really interesting, Jason, you had talked about, I don't know how much of this was in the podcast. Like they tried to contact a lot of people. And people don't like, do not want to have contact with the police. Like that's just like, you know, a cop sends you a postcard. Are you gonna be like, yes. I do want to make an appointment to talk to you, cop, from California about something I don't know anything about. They were cold calling, I'm asking them for blood samples. If you can imagine, like. Not a feeling. Yeah, getting that phone call. But that was before the technology was as some of the techniques hadn't advanced as much as they had today and that search took 10,000 hours in all these phone calls and they had to get additional tests to sort of keep seeing if they were going in the right direction down the family trees. Barbara did the same search of. For the same person. A few years later, for the same person took 10 hours and none of the extra phone calls. The databases are growing, the techniques are getting more advanced. I think there are some, a couple of different studies have shown that if you're a white, you know, European descent and you live in America, you're right. That's 60% right now and it's gonna be like 90% in a couple of years out of third expecting, right? Exactly. So like people like me, we are like already exposed fully. You found a thousand relatives when you did your 23 and me. Including your cousin, David. Can I respond to something that John said really clear before we move to the next question? So I think thinking about this and saying there's no victim or who can't identify a victim is not exactly the right way to think about it for a couple of reasons. One is even in the Golden State killer, they actually ended up, I think with two wrong hits before they got to the right person. And so there could be collateral damage on the way. And there was a case where this technique was used and they actually misidentified somebody who was arrested made bail. So I'm not saying he was actually detained for the 45 days but was in fear of law enforcement prosecution until they realized they didn't have the right person because they are looking at familial matches and it is still an emerging technology. And then the other thing is in terms of constitutional rights, we don't, I think we gotta take a step back sometimes and not just think about the victim. Because I think if we think about it in terms of how close we are to the victim, yes I think we would wanna sacrifice privacy for the desires to solve the crime. But if we think about every constitutional right that way, then why wouldn't we want to throw out all process? Because it could potentially get us closer to finding the killer if it's your sister or your mom. I get that, so that inclination I understand but I don't think this is victimless because I think society suffers. I think society is a victim when we lose privacy because one of the beautiful things about America is that when you vote, nobody knows who you vote for. And that's one of the things that protects your right to vote. It's a form of disenfranchisement if the state started to surveil your right to vote or your ability to move about your home, to form intimate relationships. All those things would potentially fall I think once we start looking at eliminating privacy. And I think with the individual not being contacted, you just find out your name, you don't even have to necessarily contact them. I think, and this isn't necessarily constitutional rights issue, but I think one thing I'm thinking about is just improving digital literacy. So when individuals do upload this, I would have less of a problem if they'd had a conversation with their family saying, hey, I'm gonna upload this information from 23 Me, just to let you know. I really wanna find my other family members like what do you guys think about this? And like so many things with technology right now, child's safety is another issue. There's so many issues where I think we need to have more conversations and more digital literacy. And I would have way less of a problem if individuals had conversations with their family and said, by the way, you know, your DNA is linked to mine, unlike fingerprints, unlike a lot of other information. How do you feel about me exposing this to the world? And like John, I try to keep a low profile. And with the Facebook example you were giving, I mean, one thing I always tell people is like, you can definitely take pictures of my baby, for example, but I don't want that on Facebook, right? I think it's kind of a similar conversation where you know the repercussions of something and you don't want it to be exposed to the world. Well, facial recognition is the analog there because we know that there's been a lot of progress made in facial recognition technology because of social networks. And I think about this too, and we talk about having the conversation with people who are related to, but people are related to like one million and one of them have not been born yet. And you think about the acceleration of this, and it's very easy I think in these conversations to say, well, I'll be dead in 100 years, but you know, that's why I think that the guardrail conversation is an important one to have. Agreed. Conceptually, I agree with Nilla in almost everything that she said. However, there's a couple of nuances that I think change how I feel about what she said. In the Golden State Killer case, it wasn't as if it were a bad DNA hit. Mistakes are made because people end up narrowing to maybe three suspects and they start following the wrong one. That is not unique to DNA. That's a police work thing. You have to figure out who the right guy did, did this crime. Once they identified the person they thought was the correct person, because you make mistakes along the way, the DNA wasn't an incorrect hit. It was a direct match for the guy that was Golden State Killer. So the mistakes that were made were not at all linked to any DNA issue that we need to worry about guardrails. Those are the generic human errors that are made during the course of the investigation. Two other problems. Exactly. As it comes to privacy going away, I completely agree with the concerns about privacy going away. But remember, privacy going away is something we all should stand up and link arms and just holler about and say no to if it's the government potentially doing that to us. This is not the government. This is individuals doing exactly what Nilla said. They're not checking with the family members and even if they were, how do you legislate that? We pass a law that says you gotta talk to your family members. I mean, that's not a policy that you can follow through on. So the point is, this is something done individually at the granular level by people that make this decision in a private setting. Government has nothing to do with that. Right now. Right. Right, Jason? Even right now, it's when you talk about if, but I'm not denying, of course, I can imagine many things that can go wrong are people that can do bad things. So I'm not suggesting that. But I am saying, we need to figure out really what the problems are and identify those bad scenarios. There are many good ones, but we have to figure out what those bad scenarios are and I would suggest that many of the really bad ones don't have police underneath them. It's other things that can be done with this information that have nothing to do necessarily with police. I think we have time for one or two more questions. That one in the back? Hello. I'm gonna take the question to a slightly different place because you've been speaking mostly about using DNA matching, you know, database to database matching. But of course, DNA can also be used for phenotyping to predict what a victim looks like or what a perpetrator looks like. I'm proud actually to be the person behind one of the programs used in a minor way to catch the Golden State Killer because we predicted something that helped law enforcement in that case reduce the suspect list or at least prioritize the suspect list. I'd be, and that's the most popular program around, which many of you probably in this room have used, which is Prometheus. I'm curious what guardrails your panel, you as panelists see, should be in place for phenotyping, predicting what a victim or a perpetrator looks like, eye color, hair color, skin color, might be prone to in the way of diseases or any other physical characteristics that can be predicted from their DNA, completely aside from DNA matching. Well, Jason, that's where Parabon began, right? Was the, yeah. Yeah, and they were, so if anyone's not familiar, this is a technique where you can basically take a DNA sample and from the characteristics in the DNA sample, come up with an estimation of what that person might look like. And that was sort of Parabon's game first. And then they had some success. And then once the genetic genealogy technique took off, they realized, oh, we have this database of really detailed DNA samples that we could just port right over to the genetic genealogy technique. And so that's partly why they've been cranking out like a case a week in terms of getting arrests or identifying victims. But in terms of- I'd love for you to address the fluke guard real question, though, because in our call, when we were prepping for this conversation, you talked about a government that is taking this technology, perhaps, and doing something different with it. And I think it speaks to your question and it sort of speaks to the idea of individuals versus entities controlling this. Yeah, so you probably saw that, in your time story from a couple weeks ago, that under the guise of free health checks, the Chinese government was collecting DNA and purportedly creating a network through which they could use autosomal tests to track the identity of dissidents. And this is in a region of the country where they have reeducation camps and so on and so forth. So it's not hard to find the sort of dystopian views of this. I think to the phenotype in question and the physical characteristics, in some ways that's an even stickier question because you then, potential for human bias to enter into that equation, to me, seems even greater when you think about unconscious bias and racial profiling. And I think you have to be really careful with that. I don't have a good answer for what the guard rules are. They are only to say that I think that one's a really sticky one to try to figure out if you guys probably have better ideas than I do. I mean, I know that it's been used, I saw a story today actually about the comparison of found missing people versus their age progression, like phenotyping profiles, it's stunning. It is really stunning. I mean, there were recreations done in the Bear Brook case for the four victims that were found in the barrels, right? Sort of like what they probably looked like. Not through phenotyping. Not through phenotyping, okay. Yeah, but it is fascinating. Well, I'm not in a position to suggest specific guardrails to the gentleman's question back there. That's a matter for people that understand that's probably far better than I. But I will say that it is that type of information that could be abused. Because if we can know hair color and eye color and perhaps we can know if somebody has a greater likelihood to be sick, or maybe they'd be more expensive to be insured. Maybe we don't want to employ them. So there could be some things like that that I think we have to be careful about. The more detailed information that's available, the more somebody can use it to make decisions. And so that's a big policy question and certainly beyond my scope. But I certainly see that as a point of reference and that's what I was suggesting when I talked about non-police guardrails. I think we have time for one more question. Got one right here. Oh, hi Russ. So I wanted to politely push back a little bit on something, Nila, you said a few minutes ago. And you said that one problem with this evidence is it's going to shine a light on people who are not related. It's gonna have bad consequences for people who are not actually at fault. Like Rabia, I come from these issues kind of from a left-ish, right-focused perspective. But I have to, I guess when I heard you say that and thought about it, I thought, okay, but compared to what? So right now we have people who face attention and are even arrested and tried and sometimes often convicted who weren't the criminal. And we know that without this kind of hard evidence, we know that eyewitness testimony is notoriously problematic. We know that jurors too often rely on hunches rather than beyond a reasonable doubt evidence. With apologies to Officer Fitzgerald, we know that sometimes law enforcement develops a theory and then works to fit the facts into that narrative. Those are all kind of very subjective facts. And I guess I wonder a move toward objective facts, maybe you have an uptake of several percent of people who are appointed at by the DNA, but on the other side, wouldn't we also maybe benefit from lots of that going away and maybe on the whole we'd have fewer folks spending years in prison before they're exonerated and so on? Yeah, I actually agree with you. I think that technology should be embraced. I think that DNA technology, I mean, it can be used to convict people. It can also be used to exonerate people. So we should definitely consider that. And I'm not against the technology by any means. I think one thing to consider is that in all of these cases, unlike sort of the exigent circumstances that we have around the Fourth Amendment where police have to rush in and make a split second decision about entering homes or collecting evidence, in all of these cases, I feel like there's a moment, there's a pause that's available here because at this point, some DNA has been collected and the question becomes uploading it to a database. In my mind, there's enough time there that there could be a third party who law enforcement consults with, probably a judge or a magistrate of some sort just to say, this is what I'm contemplating. This is the offense. What do you think? The same way we do with warrants. And I think that would work great. And I agree with you. We would probably be in a better place because humans are so fallible. And this technology, I will say though that there is human intervention at almost every step, even with DNA testing. So there can be mistakes made. So the one thing I'll say is, I think this will make things better, but I think we also have to be careful that just because this is DNA, we don't make sure that the labs are testing this correctly or that things are being uploaded in an appropriate manner. The right boxes are being checked. There's a second person making sure the right boxes are being checked because human error is still a problem. And I think we see the same thing with AI. When people see that technology or AI's involved, they often say, well, this is wonderful. Like it's better than human beings. And we see that with bias with risk assessments, which you might be familiar with like in the bail hearing context. I think once technology enters the equation, it's easier for human beings to say, well, technology told me that this is what I should do. So I worry a little bit that DNA is similar. I just want us to scrutinize and be cognizant of the issues as we go through them. But I'm with you. I think I agree with John and probably all of us on the panel agree that this technology is really powerful and the conclusion or the goal here is not to abandon it by any means. One thing, excellent comment because we've had this kind of green room discussion about that type of thing. We all know how flawed witness ID is, right? And without good science, we're relegated to using the same old fallible things that we've been using since we've had humans. Remember this though. Nila referenced a third party like a judiciary standard where you run the facts by a magistrate and say, what do you think? So that being like a cell phone, the probable cause standard is real. I mean, check and balance, I'm in favor of. But probable cause doesn't usually work at all in these things because we just don't know anything. We have no idea if there's gonna be any hits or anything from that database. We don't have probable cause to believe that that database has some sort of match in there. We're just praying and hoping. So we don't know anything. So we have no reason to believe it's there but we're gonna look because we're at our wit's end. If we had an idea who he was or where he might live, we might have some cause. But many times we have zero articulable information to actually begin to write an affidavit. How would you feel about a probable cause step after you've done the search? Probable cause to use the data? Or it's gendered kind of like wire traps. Here's my thought on that. I'm certain, I'm certain, I would never. No pressure, Chief of Police, John Fitzgerald. Yes, I would be, right now my reflex is I would be against that because if in fact, right, police generally can walk up and say hi to anybody they want to. People don't have to talk to us. We don't need a search warrant. We don't need consent. We can say, hi, how are you? Mind if I talk? No, go pack sand. Okay, or sure, I'll talk. It's all consensual. We can do whatever the citizens can do as long as we're not using our police power to intrude on privacy or freedom. We can chit chat. We can do whatever other people can do. By loading this to a database, we're doing just what everybody else can do. So, if you then say, but then if you learn something, then you have to have probably, no, no. It's an adversarial system. I agree with Nilla that what we need to do is be careful. Don't just roll over and accept it. Challenge the lab tech. Challenge the machine. Challenge the calibration. Make sure that DNA was where they said it was. Maybe it is DNA and it matches this guy, but maybe he didn't do the crime. Maybe there's another reason why it's there. There's lots of things to challenge it. But just to be able to use what's already out there to be able to identify, because remember the expectation of privacy that you're imputing is the suspects. He has none here. I am not interested in giving him some that he doesn't have under the Constitution. And yet, you're not on Facebook yourself. Well, thank you very much. I think we're going to be hanging out for a little bit to answer your questions up close and in person for a few minutes. And I really appreciate everybody coming out of this conversation. I could do this for four more hours. I don't know about you guys. This is like not a third time we've had the conversation. It's been different each time. It's been fun. So thank you very much for coming out. We really appreciate it.