
Loading summary
A
Hey, everyone. I'm Ann Emerson and this is Criminally Obsessed. We see your questions about the search for Savannah Guthrie's mom, Nancy, and I am right there with you following any potential leads and theories on this case. The wrench attack theory is gaining traction, but what exactly is a wrench attack? And we've also heard that there are people who can and have de masked the suspect in her doorbell video. Is that even possible? I decided to speak to someone who could give us some answers.
B
And I'm telling you, if that was the case, I would done it. I would have done it on that first day when the video came out. The FBI would have done it.
A
Morgan Wright is a go to expert for Internet investigations at the FBI. He's been tracking the Nancy Guthrie development since day one. And he firmly believes this is not just a missing person's case. It's a targeted attack. He says it's a no body homicide. And he believes there are critical steps that law enforcement must do in order to solve this case. The same tactics used to find the Gilgo beach serial killer.
B
I don't want to say they're going to stall, but it's going to be kind of like we don't have any forward momentum anymore. So how do we jumpstart this thing?
A
He's got some great insight on the Guthrie investigation. Be sure to like and subscribe so you don't miss any of the updates. These stories that we're all obsessed with. Now let's get into it. It's great to have your expertise on this right now. Now you're the founder of and the CEO of the national center for open and unsolved cases. Tell me, tell me what you work on.
B
So we are a DOJ funded nonprofit. This actually came out of work, believe it or not. Started years ago out of the intelligence community. It's the use of AI. It's the use of technology. Structured prompts and criminal investigations. The use of what we call ad tech. All the mobile data that's collected in addition to your cell site investigative genetic genealogy workstation. So it's called the analyst workstation. So that's what we do. So we work with law enforcement. Assumptions are killers. I work on hypothesis. It's like a science thing, right? You test it. It sounds like I'm being a geek, but look, I used to carry a gun to Gary badge. So in the Guthrie case, the problem there is that when one theory doesn't work, we come up with another theory. Now it's the wrench attacks. Oh, she could have Been the victim of these wrench attacks.
A
What are wrench attacks?
B
So wrench attacks are, you know, we used to call it rubber hose cryptanalysis, right. If you can't get the password out of somebody, you beat them with a rubber hose, right? Then you get their passwords. They get this because they find these crypto wallets that are exposed on the dark web, or that information's out there, it points to particular person. Then they. Somebody send somebody physically out there to use physical influence to do it. That's what they call. So basically, it's like you use a wrench, a real wrench. Like you use a wrench to get the password.
A
Got it.
B
You got Tony Sopranos guys going out there going, hey, you know, and what do you think?
A
Do you think this is a wrench attack?
B
No. Early on, I said, and here's the problem. People said, it's a burglary gone wrong. I said, well, either it's a burglary gone wrong or it was never a burglary. And for a burglary to gone wrong, you have to believe that a person that went there to steal stuff from inside the house, once confronted with somebody inside the house, said it would be less risky to abduct that person and flee than it would be to just flee. So what? So there's. There's very. There's a lot of things that. That contraindicate that it's a burglary gone wrong. I haven't found anything yet to contraindicate that it's a targeted abduction. And I'll give you one piece. Why? It goes back to the video. Everybody saw the intruder on the porch for flashing the. The, you know, the. The plant material and stuff. But he is comfortable being on that porch. And this, what I goes back to the serial crime profiling. You got to look at human behavior. He's comfortable standing on that porch. He's not worried about being identified. The ring camera was taken to avoid identifying a vehicle because a vehicle is very identifiable. It's not like fish. It's not like putting a mask on and defeated facial recognition. So that's why I haven't seen anything yet to indicate that it was anything other than a targeted abduction and most likely gone wrong because she's 84 years old, compromised cardiac, violent encounter at 2 in the morning that wasn't survivable.
A
Is AI technology software there to be able to start looking at faces that are obscured?
B
No, and here's why. It's called information theory. You can't analyze what you never collected if you Put on a mask, like what the intruder is wearing, and I take a picture of you. AI can't analyze what it doesn't have.
A
Tell me, why wouldn't a de masking technology work?
B
There's a case actually came out of Germany. Interpol worked on this guy. This was a German guy, had been distributing child pornography, I think it was. But what he'd done is he swirled his face. There's those effects on there where you can like swirl your face. Well, it didn't delete any of the pixels. So when we look at each other or we do a photograph, you've got your. You have pixels. All the German authorities did was they did an algorithm that said, well, let's just reverse that. So they reversed it and put the pixels back in their original place and they got the guy's picture. The reason they could do that was because none of the pixels were destroyed or covered up. Take that on the other hand, though, is that if I take an editing program and I erase the pixels, take a picture of you, but I erase your face and I erase the pixels, then I take a screenshot of it, so you can't even recover it from the original image. You're not going to recover your face. There's nothing there to recover. The other thing too, Anne, is nest cameras, ring cameras, they compress their video significantly because it's about bandwidth, it's about storage. So you're not getting the. You're not getting a high resolution picture as you would if you'd used your iPhone or Android to do it.
A
When do you think or will this technology get there?
B
No, again, it goes back to, you can't analyze what you never collected. I mean, what happens is now there was a program, and I can't think of the name, but it's where somebody said, hey, they took a blurry picture, for example, of Barack Obama. They took a blurry picture of them, they ran it through and they said, okay, unblur it. Well, it turned him into a white guy. It did it by scanning all of these images it had and said, oh, we think this is the best guess. It did the same thing with Lucy Liu and Alexandria Ocasio Cortez. I mean, those were the high, high profile examples. And I'm telling you, if that was the case, I would have done it. I would have done it on that first day when the video came out. The FBI would have done it.
A
Well, tell me, Morgan. What? How can new technology really help this case? Using the evidence that we have so
B
far right now, you know I think we. There's a lot of things. Look, even though we can't get the face out of this picture, there's a lot of evidence we can pull out of this picture. For example, if there was enough video, you could get gait analysis, G A I T gait analysis. So we could start getting an idea of how this person walks. You can actually do. It's not scientific yet, but I can narrow down people. I mean, you see some people walk differently. Some people have a skip or a hop. So you might be able to narrow it down. We can the all of the features when they went out and measured everything. So you can do things called photogrammetry. So I know that those blocks in the walkway there and everything, those are a fixed length. You go out and measure those. Now I can compute with extreme accuracy how tall this person is, how big of a shoe is this person wearing, and most likely how much does this person weigh? Well, then what do I do? I start going to things like my DMV or other stuff said, hey, I'm looking for all males that are five, ten that weigh. I'm just making this up, folks. I don't know if this is the right number, but 5, 10, 165. I can now start getting a scope to do that. It could also potentially do enhancement of things that are in that image that we can't see very well. So it could do that. The other thing we're looking for, it's like in the Kohlberger case, when he was driving into the crime scene, he shut his phone off, committed the murders, came back out, turned his phone back on. Well, that is an abnormal behavior because I don't drive down the road and then just randomly decide to shut off my phone. So there's two things you get there. Their cell site, location information, and then there's this thing called AdTech. Every app you put on your phone provides data. That data goes back to the owner of the app. They sell it to aggregators. Perfectly legal. Those are the tools. I think that where we're going to get the biggest breakthrough. There's so many things about this case. The problem is, though, ann, humans have two types of thinking. Type 1, or System 1 and System 2. System 1 is built on heuristics. Give me a shortcut. I need the story to make sense. You've been in the news business for a long time. People help me make sense out of this. System two is the tougher thinking. That's where you have to take longer. It's more of a cognitive Load. Sometimes people want, what's the punchline? Give me the punchline. Well, you can't get to the punchline yet. Sometimes the answer is I don't know. So I think that's kind of, kind of bring it all back in. That's kind of where we're at. There's a lot of unknowns in this Guthrie case.
A
I get a lot of questions on our show about the ransom notes. Do you think that these were total red herrings or do you believe that there was a kidnapper, an abductor, who actually used a ransom note at all?
B
I don't have any inside information, so I'm just. I'm going luck with you. But I think that all of these ransom communications, they were. There's a term for it, but it's called parasitic communications. I haven't seen anything that rise to an evidentiary level to say, yes, this is a valid ransom communication. Because if you really want to do a ransom, the last thing you do is you blast it out to tmz, which ensures that the entire world's going to know about it. You don't want, you know, most of these things. You don't want law enforcement involved. You don't want this. But they blasted it out to TMZ. Somebody got their 15 minutes of fame. For it to be a valid ransom communication, you have to have normally two things. You have to have a demand, which they did, but you have to have a way of contacting, simply doing a one way communication, blasting it out, with no way to get back. That makes it very difficult to engage in negotiations. This whole thing about the cartel being involved, this is not their model. Their model is they like express kidnappings. I take you. I can take you somewhere, go to an atm, get money, or I want your money.
A
I.
B
They're not into this long term. Let's negotiate, let's get visibility. So with the ransom communications, all I would say is, I haven't seen anything yet that's publicly available that gives any credibility to these communications being other than opportunistic and parasitic.
A
Do you have a theory on the Guthrie case right now?
B
No. I mean, again, and I don't want to. I have. I have a hypothesis. And so if they test that, you can test it wrong. But for me, my main working hypothesis now, it was a targeted abduction, but it was a targeted abduction that went wrong for a variety of reasons, right? 84 years old, needs medication, violent confrontation. We know she's bleeding. When I was down there, I said, you got to say the quiet part. Out loud, which is, you need to work this as a nobody homicide. You can't work this as a. As a. It's no longer a missing person, not in her condition, because it fundamentally changes the message you give to them. So that's. And then. But the question comes in, so where's Nancy? Right. So I actually went out and walked that area. Because anytime you do nobody homicides, you're also looking for clandestine graves. So you got three types of grave sites. Basically, open grave sites, clandestine graves, or concealed. Like I put a body in a building or something. Well, nobody's. Nobody's taking the time to stop to dig a hole, because that is hard pack out there in Tucson, in the hills there. So the question is, was it concealed? Was it open? If it was open, most likely, I think somebody would have found it by now.
A
Why do people believe that this investigation is at a standstill right now?
B
The time, I think from a forensic standpoint, anytime you have a hair that doesn't have a root on it, then the ability to do DNA analysis on it and put it into CODIS, the Combined DNA Index System. So to put it into CODIS, they have to have these things called 20 Short Tandem Repeat Markers, 20 STRs, 13 for a partial. If they don't have that, you can't really put it into CODIS. If you have below 13, you can't put it into CODIS because then you end up matching tons of things. So it's really not very good. Parabon, which actually is in the county over here next to me over in Fairfax county, they do this composite snapshot so they can take DNA and recreate a face out of this. This actually, this actually happened in the Brittany Phillips case out of Tulsa, Oklahoma. I did that case last week on my crime reconstructed podcast. But they had DNA. Didn't test it for 15 years, but they sent it to Parabon. Now, did it match a guy? Yeah. The only problem was the guy was with one of her girlfriends in the other bedroom. That's where the DNA, the semen and the blood came from, not from her attacker. So for 16 years, they were going, oh, this DNA, this semen belongs to the attacker. It didn't. It belonged to the boyfriend of her friend that were in another bedroom. But parabol. But the, the, the composites they can create out of these snapshots, what they call them. Eerie. So, I mean, but we are still a ways away from that, because if the DNA is convoluted, then you have to deconvolute it. And even that stuff hasn't been approved for use in federal court yet. It doesn't meet the evidentiary standards.
A
What do you think is next for
B
this case you were just talking about, I think, right when we started, or you're talking about one of the cases you're going to cover is Rex Sherman, Gilgo Beach. Right. I was. We were actually this. I was doing a. Once a month, I do a thing called Saturday Ramp. But we were talking about cases, and I talked about this case, and it was basically what solved. What solved us this time. People said DNA solved it. I said, no, DNA didn't solve it. The creation of a task force solved it. That went back to the original lead that found the Chevy Avalanche. That led to the identification of Rex Shearman. That led to then getting the DNA sample. That's what solved the case. Is a task force approach, a true task force approach. It's like working a cold case. I've got one down here that's this thick, and it's like. It takes a long time to go back and look through this stuff. But what you do is, I don't want any. I don't want to talk to any of the prior investigators. Come in, fresh eyes, fresh face, no preconceived notions about what you believe or what you think. No such thing as a smoking gun. But I think a task force approach, bringing this stuff back together and start from square one. A lot of smart people in the original investigation, but because it was fragmented, that lead on the Chevy Avalanche got taken and it got buried. It never got followed up. When they did the task force, they went back to square one. Somebody followed through on that Chevy Avalanche, and now we've got sentencing for Rex Shearman as a result. DNA identified Rex Shearman. But what. What really broke the case open was the lead on the Chevy Avalanche.
A
I completely agree with you.
B
Everybody says, well, it's because of Savannah. It's because nobody knows. We don't know. And there's only one person that knows right now, and it's the person who committed this offense. But what we can look from that is go. If Nancy passed away, you know, she died shortly after this. You have no leverage. You have no targeted abduction anymore. You have a body problem. So the risk of driving around with somebody in your vehicle increases. It's like serial killers. It's called geographic profiling. You get this thing called distance decay. They only go out so far where they say, oh, the risk is too great to take this body out anymore. This is as far as I go. And from that, you can start figuring out, okay, well, where's their circle of life, their home, their work? So in this case, I'd look at. You're going to be looking at a major thoroughfare. Somebody might say, well, it's only like, an hour to get to the Mexican border. That's true. But what you do is you run the risk coming across the border of having a body in your car. So was it done before that? Then you start looking at. There's a lot of studies done and that show with people who dump bodies off interstates or off roads, how far out do they go? It's usually about 3/4 of mile is kind of that distance decay, too. So I would start looking at major roads, and I'd go 3/4 up to a mile in each direction on each side of it, and I'd start kind of doing that as a grid search on all of these potential egress routes out of there. To say if somebody was going to do that, they don't want they. If they get stopped, they don't want to have a dead body in their car. So what do you do? You get rid of it as soon as possible, Drive off a road somewhere, Conceal it, and then get back on the road.
A
With all of the people you've spoken to on this case already, have you. Do you feel like that's what law enforcement has done? Have they done that grid search? That's right off of. Off of the guthrie property? Three quarters of a mile.
B
Yeah. Actually, When I was flying down to tucson, One of my friends and I, we developed the requirements For a bluetooth scanner with an omnidirectional antenna. We built it onto a drone and so you could scan for the low bluetooth signal. But so the. And the FBI started doing it, too. Actually, the team there was working with the FBI directly to do that. So they've searched thoroughly the area around Nancy. What I'm saying is that you look at. There's five. Five ways to get out of that neighborhood. When I was doing my analysis, Five major ways to get out of it. I would take those five major ro. Once you get to, like, five miles out, and I would look at those routes going in all those directions, and I would go up to a mile on each side. Look where the neighborhood is. It looks like a bowl of spaghetti. Somebody dumped a bowl of spaghetti. All those roads. But there's only five major ways out to get there, to get to major arteries. That's where I would start doing my work, is I'd start saying, okay, let's take this radius. Let's start taking this route, this route, you know, divide them into routes 1, 2, 3, 4 and 5 and start working grids. I don't know that they've done that. I think they've done a comprehensive search around the property, which they should. I just don't know how far out they've gone. I haven't seen anything to indicate that they are identifying this window of opportunity, these egress routes, and doing grid searches based off of an interstate or a major road or a major thoroughfare. A lot of people say, do you think it'll ever be solved? And that's one of the cases we don't know. Some cases wait 30 years before you get a breakthrough on DNA or a confession or a deathbed confession. But I'll tell you last final thought. People say, what do you think will open this case? I don't know that it will. But I would put my money on an arrest for a completely unrelated offense that somebody's about to get jammed up on big time. You just committed a robbery. You're going to prison. You're a two time felon. You're looking at 30, 40 years this time, federal time. Somebody's going to go, well, you know, I know something. I'm willing to make a deal.
A
I got something, I got something for you.
B
Quid pro quo, you know. Okay, in the words of Hannibal Lecter, quid pro quo, Regent Starling. So that's a little creepy. But, but, but, but that's, that's to me, that's the way I broke a lot of my cases. You get somebody jammed up and I'd simply ask, okay, look, you're jammed up here, but what do you know about something else? You give me something better than what I got on you, I'll take it to the District Attorney.
A
Drop a comment below and tell me, do you think a task force needs to be assigned to the Nancy Guthrie case in order to solve it? I want to hear your thoughts. And be sure to like and subscribe to Criminally Obsessed. And we will keep on updating this story with more new information.
Criminally Obsessed
Host: Anne Emerson
Episode: Guthrie Investigation: FBI Tech Expert Answers Our Burning Questions
Date: June 17, 2026
This episode dives into the continuing investigation into the disappearance of Nancy Guthrie, the mother of Savannah Guthrie. Host Anne Emerson interviews Morgan Wright, an FBI technology expert and CEO of the National Center for Open and Unsolved Cases. They address listener questions, explore technological and investigative challenges, and provide new insights into theories, evidence, and the potential path forward in this high-profile case.
Targeted Attack vs. Burglary Gone Wrong
“I haven’t found anything yet to contraindicate that it’s a targeted abduction.”
– Morgan Wright [03:56]
“If a burglary goes wrong, it would be less risky to just flee than abduct someone. There’s a lot that contraindicates it was a burglary gone wrong.”
– Morgan Wright [03:00]
Behavioral Cues from Doorbell Video
“He’s comfortable being on that porch. He’s not worried about being identified… That’s why I haven’t seen anything yet to indicate it was anything other than a targeted abduction.”
– Morgan Wright [03:56]
Limitations of Modern AI and Image Processing
“AI can’t analyze what it doesn’t have.”
– Morgan Wright [04:21]
“If I erase your face pixels and take a screenshot, you’re not going to recover your face. There's nothing there to recover.”
– Morgan Wright [04:52]
Prospects for Future Technology
“You can’t analyze what you never collected... They ran a blurry Obama photo but the unblur program just made a guess—not real data.”
– Morgan Wright [05:54]
Gait Analysis & Photogrammetry
“If there was enough video, you could get gait analysis… Then you start going to things like DMV… I can start getting a scope.”
– Morgan Wright [06:38]
Digital Forensics: Cell Sites & AdTech Data
“Every app you put on your phone provides data. That data goes back to the owner of the app. They sell it… Those are the tools where we’re going to get the biggest breakthrough.”
– Morgan Wright [07:29]
Cognitive Mistakes in Investigations
“Sometimes people want, ‘What’s the punchline?’ …Well, you can’t get to the punchline yet. Sometimes the answer is ‘I don’t know.’”
– Morgan Wright [08:34]
“Haven’t seen anything that rises to an evidentiary level to say, yes, this is a valid ransom communication… For it to be valid, you need a demand and a way to contact back.”
– Morgan Wright [09:27]
Moving from Missing Person to Homicide
“Say the quiet part out loud, which is, you need to work this as a no body homicide. Not a missing person, not in her condition.”
– Morgan Wright [10:36]
DNA and Forensics Limitations
“If you have below 13 [STR markers], you can’t put it into CODIS… Parabon can create composites, but if DNA is convoluted, you have to deconvolute it.”
– Morgan Wright [12:05]
Learning from Gilgo Beach Serial Killer Case
“The creation of a task force solved it… It wasn’t DNA, it was the Chevy Avalanche lead.”
– Morgan Wright [13:15]
Grid Search Strategy
“I’d start doing that as a grid search on all of these potential egress routes out of there.”
– Morgan Wright [15:39]
Likelihood of Resolution
“I’d put my money on an arrest for a completely unrelated offense… and then someone says, 'I know something, I’m willing to make a deal.'”
– Morgan Wright [17:56]
“Assumptions are killers. I work on hypothesis. It’s like a science thing, right? You test it.”
– Morgan Wright [01:43]
“Quid pro quo, you know. Okay, in the words of Hannibal Lecter, quid pro quo, Agent Starling.”
– Morgan Wright [18:06]
(Shows his dry sense of humor and deep experience with casework)
“With the ransom communications, all I would say is, I haven’t seen anything yet that’s publicly available that gives any credibility.”
– Morgan Wright [10:17]
“Some cases wait 30 years before you get a breakthrough on DNA or a confession or a deathbed confession.”
– Morgan Wright [17:23]
This episode sheds light on the technological and procedural realities of the Nancy Guthrie investigation, debunks widespread public theories, and emphasizes the importance of fresh investigative strategies like a task force and systematic grid searches. With a sharp, direct style and decades of experience, both Anne Emerson and Morgan Wright deliver a sobering, knowledgeable analysis. The consensus: a resolution will take rigorous evidence-based work, perseverance, and perhaps a lucky break from an unrelated arrest.