GenAI for Private Investigators

Today, we're talking about something that's changing the investigative landscape: Generative AI. It's everywhere, it's powerful, and it's already being used by private investigators — but not always wisely. 

Dean Beers joins Hal to talk about how PIs can leverage GenAI ethically and effectively, where its limitations are, and why understanding this technology isn't optional anymore. 

Get Dean's class: 

GenAI for Private Investigators: Applications, Limitations, & Ethical Considerations

Host: Hal Humphreys 

Guest: Dean Beers

Music provided by Jason White, who composed our theme.

Special thanks to Kim Green, who produced this episode.

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Pursuit Magazine and PI Education are part of StoryboardEMP, a media and education company in Nashville, TN. Pursuit, a free online magazine for private investigators, explores all things investigative, from gumshoe techniques and surveillance tech to industry news and crime in media. PI Education, its sister brand, provides online continuing education for licensed PIs. Pursuit and PIed are owned and edited by husband-wife team Hal Humphreys, a PI, and Kim Green, a writer and radio producer.

In this podcast, you'll find episodes that dive deep into the work and the business of private investigations. And at PI Education's YouTube channel, you can dive even deeper into the knowledge pool of this fascinating profession, with regular briefings and webinars. Subscribe to stay up to date!

Pursuit Magazine: https://pursuitmag.com/​
PI Education: https://pieducation.com/​
Our channel: https://www.youtube.com/c/PIEducation​

Transcript

SPEAKER_01 0:02

Welcome to the Sound of Pursuit. I'm Hal Humphreys, your host, and today we step into the bright and blinking carnival that is generative AI. It promises efficiency, insight, maybe even omniscience. It also promises to make fools of the unwary. The machines have entered the investigative trade, and some of them are already wearing trench coats. Dean Beers joins us to separate the tool from the toy. We talk about how private investigators can use generative AI without surrendering their judgment, their ethics, or their good name. We examine the limits, the liabilities, and the quiet danger of trusting a confident algorithm. This is not a conversation about the future, it is about now. And whether we intend to lead it or be led around by it. Stick around.

SPEAKER_00 0:52

To get more intel on the real work of private eyes, subscribe to our YouTube channel, readpursuit mag.com, and take classes at PI Education.

SPEAKER_01 1:01

I'm here today with Dean Beers. Dean's been a fixture in the PI world for over 35 years. He's a licensed private investigator, an instructor, and a course author who's covered everything from death investigations to legal investigations. But today, we're talking about something different, something that's rapidly reshaping how investigators well, hell, rapidly reshaping how everybody works. Generative AI. Dean has just launched a new course for us at PIEucation.com, Gen AI for private investigators, applications, limitations, and ethical considerations. It's a foundational course on how AI tools like ChatGPT and others can enhance investigations where they fall short, and the ethical guardrails every PI needs to understand. Whether you're already experimenting with AI or you're skeptical about the hype, this conversation is going to give you clarity on what's real, what's risky, and what you need to know. Before we get started, Dean, I want to mention something important. I know you had some health scares not too long ago. How are you feeling these days, buddy?

SPEAKER_02 2:05

Yeah, thanks, Hal. You know, at our age, I guess that tends to happen. And even with my experience, I knew it could happen. But yeah, uh, so July 6th, uh, 2024. So, you know, not quite two years ago now, uh, my wife and I decided to surprise our daughters who went to uh Lake McConaughey in Nebraska. They never get together for a mini vacation, and we thought we'd go surprise them, and and we did. Um, the first morning we were there, everybody's lounging by the pool. Um, I felt pretty uncomfortable. I swam underwater to scare the grandkids, and uh I felt a tug on my chest. I tried to uh figure out what it might be, and eventually I realized what was going on, made it to the edge of the pool, tried to relax. Um, that didn't do it. And I got up to go to the room and get some aspirin, and uh, you know, things happened. My wife and family saw that I wasn't right. And thankfully, uh my wife, Karen, you know, she uh followed me to the hotel room just in time to uh realize what was happening, uh, asked me if I wanted her to call 911, and uh thankfully she did. Um it was a small local hospital. Um they took me there, stabilized me, but had to fly me from central Nebraska to western Nebraska uh where they tried some stints. That didn't work, so they had to drive me by ambulance because of tornadoes to the other end of Nebraska, to Lincoln, where I had uh bypass surgery, triple bypass surgery on Monday the 8th. Um went through that process and came home uh Friday uh the 12th. And um, you know, my whole family stayed there during that time and uh it changes your life. When I saw the cardiologist, uh my heart functions pretty bad. And uh she told me basically I was walking dead that whole time and surprised I even made it. So it it changes a lot of things uh for you. I did well on cardio rehab, but about eight months later things got were worse, and I ended up getting a device implant, which basically keeps my heart going uh at a at less than half the pace it should be. So uh you got to make a lot of adjustments and and to connect all this, AI has a lot to do with it in a way, but I'll look forward to talking about that. Thanks very much, Hal. And also I want to add um a lot of investigators around the world came together for me, and that included you, Kim, and Stephanie. So thank you very much. We'd never forget that.

SPEAKER_01 4:39

Hey, Dean, we're just glad you're still here, buddy. Um let's uh let's dive into the topic. Let's start with the basics. What is generative AI and why should private investors uh private investigators care about it right now?

SPEAKER_02 4:53

Yeah. Well, there's a lot of reasons to care about it, but let's we'll focus on uh the advantages for us versus the disadvantages and the warnings that that that we're all going to become familiar with. But uh generative AI basically it's uh as you kind of touched on, it's an algorithm that constantly updates and changes itself. Um we've all been using AI in some fashion over the years through algorithms, and then in recent years, we may not have even known it and started learning about it when we would do a chat online or make a phone call and say, This is your AI assistant. How can I direct your call? But now we're able to interact with AI and use it for ourselves in our daily lives and in our work. Um, so there's advantages to it, and that's what the course is about, and that's what we're going to talk about.

SPEAKER_01 5:37

Okay. Now you've been teaching investigators for a long time, Dean. How has AI changed the investigative landscape?

SPEAKER_02 5:47

So it there's different perspectives of it, but what I'm seeing finally develop, and this is part of why we did the course, is for investigators to use it in the right way. Uh, to begin with, like administrative processes, uh, helping with research, helping with reports, but also getting into some of the more investigative aspects of it. And there's some warnings that come with that, we'll talk about, but uh it but not using AI to do the investigation, but using AI to enhance it. So we really need to harness it, or we're gonna all fall behind on it, not necessarily be replaced, but our clients and colleagues are gonna be ahead of us if we don't get with it.

SPEAKER_01 6:28

Yeah, and I've noticed, you know, I I've been playing around with um several different um generative AI programs, softwares, and one of the things I've noticed is you know, if you're doing investigative research work, that kind of business, um the possibility for hallucination is very real. Um I had I had I had one case I was working on. I tried to do a little research using Chat GPT, and I got absolute made-up information back. It was very hopeful at the beginning. I'm like, wait a minute. I asked her, I said, did you make this up? And I said, Yes, sorry, I didn't mean to do that. That's against my protocols. Yeah, how did you do it? Um so yeah, there are some there are some pitfalls. Um without getting to the weeds, where could AI fit into our work as private investigators?

SPEAKER_02 7:17

So uh there's a lot of different areas. Administrative at least probably one of the key things because we can use it to create more time for ourselves. Um and that's really one of the focuses. Uh you can use it to do even market research for yourself and your agency. You can use it to do uh SEO for your websites, uh, you know, a few different things. We once we started understanding how it works and how it doesn't work, uh, we started using it to, for example, in a case, take a bunch of our reports and notes and merge them into one. Um we have cases where the client will ask us to develop questions for a deposition of somebody in the opposing side, and we can take all of our notes and ask it, and we have templates of questions and we can ask it to customize those questions, and then we can further ask it to include the answers that we've already had in our notes. So it can save quite literally hours on time, and uh, you know, and that's not lost buildable hours because there's other ways to work that out, but uh but once we get from administrative and merge that into investigative, there becomes a lot of advantages. Uh, there's investigators that will use AI first to analyze a case. I personally don't find that's an ethical thing to do, but I do find that you can use it to take a look at your case and and say, you know, how does this look in the real world? Now, again, like you said, we got to watch for hallucinations, and there's a variety of gen AI out there. Uh, and you got to really know what one works best for what, and you got to learn these prompts and things in order to ask the right questions and have that conversation with AI.

SPEAKER_01 8:55

Yeah. Um you know, one of the things I have I have begun doing this. Uh, I'll have several interviews that I've done um on a case with different witnesses. Um, and uh so I'm gonna throw a little caveat in there. Um be very careful about what you upload into the various AI engines. Some of it becomes uh out there into the ether for anyone to use, others are closed systems that you can upload stuff into your system and it it will deal with it right there. Right. We use an AI program that is is self-contained. It it you feed it the information, it leads it, reads that information, analyzes it. One of the things I've found, Dean, is if I've got uh say 10 witness interviews, I can upload those into that software and have it find common themes among the interviewees. Um and that that that helps me define hey, this thing everybody noticed, this thing one person noticed. It's it's helpful that way. Um what are what are the biggest misconceptions you hear about um artificial intelligence from investigators?

SPEAKER_02 10:07

Uh one that it's gonna replace us, um, but yeah, I think that's kind of the opposite. It but it can replace what we're doing um in in a sense. Your example is a very good example. You can use it to find uh commonalities as well as stark differences and isolate those things. Um so some of the things that investigators find that do wrong is it it can, for example, just do all the thinking for you. I think that's probably like the big, the big overall picture is that it can replace the things that we're doing and um uh in the investigative process. You can upload uh you know certain items to it, you know, with the security things you mentioned. We mentioned those in the course, what to do. Um, and it will spit out uh, you know, things that that uh answer the question. You can do a report and you can send it off to your client. And I kind of put that analogy of it's if you do it that way, it's like doing a report from one of the data providers, downloading that report, copying and pasting it to your letter head and sending it off. I think that's probably the biggest problem um that investigators and other professionals have with it. Um if you watch attorneys give webinars on AI, they're very good about giving these warnings out. Uh, but one of the other issues with AI is just the fact that uh it instead of using it constructively um for uh for advancing our day, advancing our product, uh, you know, instead using it thinking it's going to replace and not watching for those hallucinations, doing like I saw one on a blood spatter research report. And uh I have uh areas of uh of expertise within that. And I'm reading this report and it's using the wrong terminology, it's using some of the wrong uh uh uh equations to come up with answers. Um and it's a very trusted, this one particular, but I've seen others, very trusted uh expert. But unfortunately they didn't proof the what was spit out. And I think that's the other problem.

SPEAKER_01 12:11

There's no faster way to lose credibility than let you know AI do the work, sign your name to it, pretend like it's yours, then have someone find out that it's just hallucinated or not right. Um Let's talk. This is this is a good chance to talk about the risks. What should private investigators be concerned about when using AI?

SPEAKER_02 12:32

And with that little caveat I just added, you know, the risks that we need to watch, not just not including the hallucinations, false research, you know, completely made up stuff like you mentioned. But one of the things I think we need to watch for in risks is the other, the what we see in other product coming to us. So we need to watch to make sure that stuff isn't plagiarized. Uh, because you mentioned, for example, AI basically uses your product if you're not using a secure system to learn and develop to other people. Um so if we're not looking at other product coming to us from, say, opposing side and looking to see if that was plagiarized in some way, if it was AI developed, uh, I think that's an important risk we got to watch because it could uh it could definitely hurt our side. But with that is yet another caveat, and because I've seen this where people will say, is this AI developed? knowing that they completely wrote a product, like our articles or something. And AI will, I've seen AI spit back, or other people tell me it's spit back. Yes, this is AI generated. Well, that's completely false. Um, it may be AI enhanced or AI helped or whatever. Um, we use AI to proofread some of our stuff. Uh you got to watch and make sure it doesn't proofread in a foreign language, uh that you know, it's happened to us. But those are some of the risks. We just really got to be aware that other people are using it, other investigators, other attorneys. And we've seen this in the news. We just got to watch and make sure that we're not in in receipt of that. And by the way, that includes from resources that we get for our investigative research, you know, where it comes from. Is somebody just slapping out AI-generated reports to us when we're asking for something like uh, you know, uh, we already know that databases are using AI. They've used algorithms, now they're using those generative algorithms. Uh, so we got to really be careful of that. Um, I think it's going to become a big problem before it's resolved in what we see.

SPEAKER_01 14:29

Yeah, I mean, the the things that scare me are um, you know, if I'm working with a transcriptionist service, I can have them sign a non-disclosure agreement so that they they are bound not to talk about what we've discussed. Um, again, going back to if you're uploading sensitive case files into Chat GPT, be very careful about that. Um, you know, we we've still got to think about confidentiality, privacy, those kind of things.

SPEAKER_02 14:54

Um, that's in our course.

SPEAKER_01 14:56

Yeah, yeah, yeah. Um let me ask you this: when should an investigator not use AI?

SPEAKER_02 15:04

Uh my first thing is to go back is don't use it as your initial step in conducting an investigation. That's why your clients hire you, that's why they hire us. Uh if if that's how you start, like you said, you lose credibility because believe me, it's not going to take long for an attorney, you know, they're bright people. Uh, they can read a report that we send in from one of the database providers and tell that it's nothing different than their paralegal did. If we start using AI to create our investigative work product, that's where we're going to be replaced, or you, like the investigator who does that, is going to be replaced. Whereas the rest of us, you know, yourself, us, you know, we enhance it. We, you know, we may use it just for proofreading. Um, but that's that's definitely one of those things that I think uh investigators need to watch out and be careful of. They shouldn't use it to start the investigation. And on the other side of that, they shouldn't use it to finalize the investigation. If you're writing a report of uh witness statements or we write an expert report on a death, if we send this up to a secure AI and it comes back and tells us something's wrong, whatever, and we rearrange our report to do that, I'm gonna have a problem with testimony because that's no longer my report or my opinion. Um sending in a witness interview to a client, uh, maybe something, you know, AI decided that, you know, it's not practical that something could have happened. Um so it changes it. So if we don't proof all of our own work product out of uh the assistance of that we've gained assistance from AI, we're turning in something that could be uh fallible.

SPEAKER_01 16:40

Yeah, I I love the idea. You mentioned earlier the idea of like if you're gonna use artificial intelligence, put a disclaimer in there, you know, portions of this were generated or we used it for proofreading, whatever it is. I think being transparent is key. Um I also think you know, here's the thing for the investigators out there that are afraid of AI taking over our jobs. I'm here to tell you artificial intelligence is not particularly good right now at doing rural surveillance. Artificial intelligence is not really good right now at doing witness interviews. Um, I might use versions of artificial intelligence to help me nail down where a witness is located. Um to go through some database reports and give me the most likely address, those kind of things. But there is no way right now for AI to go do the interview. Um, I think a human-to-human connection is key there. There's a lot of hype about AI right now. What's what's real value versus what's just tech buzzwords?

SPEAKER_02 17:39

So one of the real values of AI is in the time saving, better work productivity. And with that, and and I'll I'll bring up the course yet again. We do explain how to take the value of AI and make it work for you. Uh, and one of those, for example, is taking what we normally charge hourly for and help AI develop things that we can chart to start charging a flat rate for. So, for example, we might use it to take a transcript of an interview, create a bullet point for a client, and then use that bullet point to create a witness statement that we're going to send to them to sign off on. Of course, like I said, we got to proof all this, make sure it didn't, you know, throw anything out there. But what would take a couple hours now takes about 15 minutes. And we still charge uh a flat rate for those services, and it increases basically our hourly uh uh rate. And our clients are doing, they're frankly, attorneys are doing the same thing. It's something that's going to become the norm. And I think that's the biggest value of it is we can go from trying to keep track of hourly rates to taking advantage of flat rates for things that have uh you know a known a known outcome uh for that. So I think that's one of the big values we have, plus the administrative side of things.

SPEAKER_01 18:59

Yeah, I love that. Taking it to help us analyze where we're spending our time, how we can realize some efficiencies. And if if you've got something that used to take three hours and you're doing it now in 15 minutes, it's not fair to the client to build them three hours for that, but it's also not fair to you to just charge 15 minutes for that. So finding a flat rate, that's a genius way to look at it. That's that's what we yeah. Let me let me ask you this what made you decide to create this course and why now?

SPEAKER_02 19:31

That's a really good question. And let's tie this all into the question that you asked me at the beginning about my health. So um in in the beginning of 2024, we started rewriting my textbook, uh uh Professional Legal Investigations. At the time it was practical legal investigations, and Karen and I were going through and writing that. And when we went to Nebraska, I brought the last chapter with me, uh, or the last part of the last chapter. And I was going to final proof it, Karen was going to final proof it, and then we were going to edit it. And then Monday, when we got back home, we planned to finalize the layout of the book. Um Frankly, that got derailed a little bit, right? But we'd already made commitments, uh, self-published. We'd already made commitments to get it out by Labor Day. And um Hal, I think you know know Karen and I well enough that when we say we're going to get something done, we get it done. Um, it almost doesn't matter what it's going to take. And I I gotta, I gotta be honest, it was not a simple task. Yeah, um, but we were done. And I'm just gonna you know show right here this is the book, um, and it's 600 pages, okay? So all I had to do was that final part, and then we we we paid somebody to take care of the cover, but we needed to come up with the back product, okay? Um, the little blurb on the back of the book there. So what I did, and it's in the course, what I did. Um, I went to AI and I said, read this file and give me a good summary of so many paragraphs, uh, three-quarters of a page, two-thirds, whatever it was, and it spit out, and I'd read it, and Karen would read it, and we'd like it and not like it. And we we did this a few times until it came out with something we liked, and then we actually totally rewrote it to us. Uh, so it's AI generated, but it's redeveloped from us. We liked that so much that we ended up deciding to do a summary for every chapter at the beginning, and that's what we did. So uh that was our first use of AI. And uh, so it took a year later because of other uh dealing with health and stuff, but I let Stephanie know hey, we're we're I gotta get this course done. We started for me a year ago. I'll get it done, and then we're gonna start this next this. Next course, and I didn't tell her what it was yet until we were done. And uh, you you, Kim, and Stephanie were pretty excited about it. We were very excited about it. Um, but I started seeing to get to the last part, I started seeing a lot of different webinars from attorneys and investigators talking about not how to use AI, not what it can do for you or how it can hurt you or anything else. Um, but instead they were just describing what makes AI work. Like it's it's this big data bank of centers. It's that it's just way too much technical information. I couldn't find a course or that really spoke to investigators on how to make it work for you. So that's why we decided to do it. And um we just decided to get it out as frankly as quick as we can to get it to you guys so we could so we could get it to our colleagues out there. So that's the connection from beginning to hear of uh how that happened.

SPEAKER_01 22:57

I love that book ending of the of the story there. That makes perfect sense. May I ask you this, Dean. Getting down to the end here, what's the one thing you would like for investigators to understand about AI after this conversation?

SPEAKER_02 23:10

We need to don't be afraid of it. Uh don't think it's going to replace you. Harness it before your colleagues and clients do get ahead of you and you do fall behind. Because that's the only way you'll be replaced, is just simply by not harnessing what you can of it. I love it. I love it.

SPEAKER_01 23:29

Get in the game, people. Um, look, I I I read a lot of newsletters every morning from around the world, and you know, there's a lot of noise about artificial intelligence. But here's the thing it is a tool that we can use. It is going to be used by other people. We may as well get in on it and learn how to use it properly. I love the fact that Dean is talking about um the ethics of using it and and the practicalities of using it. Dean, um, Stephanie, our producer, has put together this list of rapid fire questions. Are you ready for some rapid fire questions?

SPEAKER_02 24:03

Well, let's go. Al. All right.

SPEAKER_01 24:05

What's your favorite hobby?

SPEAKER_02 24:07

Uh generally uh besides, I mean, honestly, work is my favorite. It's it's a hobby and work both. But I I love firearms and history and going out and shooting. Uh it I I've I've I've been with firearms and shooting since I was about four. It actually ties into our work quite a bit, so it's it's a little bit of everything. Okay. Morning person or night owl? I've always been a morning person, especially since the coroner's office days. But uh with my health, uh, I had to become more of a morning person, uh, getting a lot done early in the morning because by by the end of a short day, I'm pretty well, pretty well done for.

SPEAKER_01 24:46

Yeah, yeah, yeah. I get that. Um, what's the most satisfying moment in your PI career?

SPEAKER_02 24:53

You know, I I really there's just so many. I mean, I I know you know this, but I think what I made the 10-year mark, um, I went out and bought my wife and I a special presents um that I thought you can only afford these if you make the 10-year mark. And honestly, that was just a huge deal. Um, I've made it a few more decades since then. I plan to make more. Um and I really think that was a landmark thing. So whenever I see other investigators say, hey, I made it 10 years, I know how that feels.

SPEAKER_01 25:26

Yeah, yeah, it is a landmark. Um, all right. So we're talking about, you know, you and I are old grizzled investigators. We're at this for a long time. What's one piece of advice you would give to someone just starting off in this business?

SPEAKER_02 25:40

Network, network in person, learn. You're we're we're always after learn and adapt and share. Uh, when we when I started out in 1987, no, there was little networking, which at the time was difficult to do, um, but nobody was willing to share how they did something. And like you, Karen and I have always been dedicated to sharing uh with other investigators, and we always will. So network, learn, and share, because we all constantly, that's what keep that's what keeps our profession alive, that that that triad of of of doing that.

SPEAKER_01 26:20

Yeah, great advice. Um, may ask you this, Dean. If artificial intelligence took over all of your admin work tomorrow, what would you do with the extra time?

SPEAKER_02 26:31

Honestly, I find another project to do on a day-to-day basis. It could be articles, courses. Um, you know, we constantly keep that, keep that going. And I we just really love doing it. So I think on a day-to-day, that would be my answer.

SPEAKER_01 26:44

I love it. Um, favorite fictional private investigator or detective?

SPEAKER_02 26:50

Uh I think uh Rockford, Jim Rockford. He he's he was my as a kid, um, but we there's a ton of them before and after him, really good ones, Mannix and things like that. Um, and and a different one for law enforcement, Columbo, who couldn't like Columbo, right? And his daughter actually became a licensed PI, uh, Peter Falk's daughter. Um so many good options out there. Yeah, yeah. But I like Rockford for the time, what he's doing then would not be ethical now, but at the time it was nothing really bad or nefarious, you know, some impersonations, fake business cards, you know. But uh it was just a fun show. Just a fun show.

SPEAKER_01 27:30

It was a lot of fun. You know, uh when I think about favorite fictional private investigator, it changes for me almost on a monthly basis, whoever I'm reading right then or whatever. But yeah, Rockford has been there from day one for me. Great, great choice. Dean, I've enjoyed this chat. It's been eye-opening. Thanks for cutting through the noise and give us a straight talk on AI. And for everyone listening, if you're curious about how to use artificial intelligence in your investigations without making a mess of it. Dean has a brand new course that walks you through some of the basics. Gen AI for private investigators, applications, limitations, and ethical considerations. You can check it out on PIEuducation.com. It'll give you three hours of continuing education. It's on demand and it is packed with real examples, ready-to-use prompts, and the kind of practical guidance you can put to work immediately. You'll learn where AI fits, where it doesn't, and how to use it responsibly. Check it out at PIEucation.com. Dean, if somebody wanted to get in touch with Dean Beers, what would they do?

SPEAKER_02 28:38

Yeah, um, you can drop us an email, associates at deathcasereview.com. And uh you could also visit uh deathfirks.com for our contact information, our phone number. Uh but yeah, we'd love to hear from anybody. Um, even ideas about courses or how we can help them out. Uh, you know, that's what we're here for. Thank you very much, Al. It's always a pleasure to talk to you. We've been longtime friends working with uh PI Education and Pursuit for a long time. Uh, thank you very much.

SPEAKER_01 29:05

Oh, we appreciate it a great deal. And for those of you listening, you've got to understand Dean can talk about all of the things he might do with his free time if AI took over all of his administrative work, but you just heard what he's gonna do. He's gonna create new courses, he's gonna write new articles, he's gonna share information with us private investigators. If we have nothing else to do today, and I don't think we do. That is Dean Beers, and I'm Hal Humphreys, and that is your Sound of Pursuit for this week.

SPEAKER_00 29:35

You've been listening to The Sound of Pursuit, a podcast by Pursuit Magazine and PI Education.