Episode Transcript
[00:00:02] Speaker A: Welcome to the Love justice podcast where we share insights behind real impact in the fight against human trafficking. Today we are diving into the role of data in fighting modern how Love justice collects, analyze and applies data from one of the world's most comprehensive human trafficking databases to sharpen strategy, improve incomes and steward donor trust. I'm joined by John Molineau, founder and CEO of Love Justice International, who is going to give us a behind the scenes look at how this data drives preventative impact, informs critical decisions and helps the team innovate for maximum effect. John, welcome to the podcast. Welcome back.
[00:00:44] Speaker B: Thank you, Hannah. Glad to be here.
[00:00:46] Speaker A: Yeah, I, you know, I'm excited for today's conversation for a couple reasons. One, one of the roles that I was initially hired for 10 years ago was a data analyst. And I've definitely gotten to see the evolution of how we our strategy behind how vision envision data being critical in the fight against human trafficking. So I think it would be really helpful, John, just to lay the, lay the foundation of this conversation and start off by just giving us a high level overview of why data is so important to Love Justice's strategy, but also why it's important in this field of anti trafficking work.
[00:01:27] Speaker B: Yeah, I mean, I mean to effectively fight anything or combat any injustice or really work in any area, you need to understand it. But it's the nature of human trafficking is that it happened in the dark.
There's not clear data points on it. If you could go in and interview and get data on all the people who are enslaved today, well, you'd be better off to just rescue them than to get data on it. And so what exists is a very, very unclear picture of human traffick.
And what is known, I think is often based on speculation or, you know, and is often not accurate. So data is desperately needed in the fight against human trafficking and it is lacking. There is shockingly little information out there about it.
You know, and you know, we intercept 2 to 3,000 potential victims every month and have for several years, or at least it's been hundreds for several years and that number has grown over the years. But we have a chance. With every potential victim we intercept who's in the process of being traffic, we have a chance to collect data about their situation, about them. We first collect data using a form called the Intercept Record form where we get information about how the intercept occurred, what were the red flags. That form is also used to guide the questioning protocol by which we determine whether to attempt to intercept someone and who was in the group and what Were the signs that noticed, how did the intercept occur, et cetera. And then we have a chance to fill out a form called a suspect form, where we get information about the alleged trafficker and demographic information, information about their identifiers and just helps us to understand trafficking and trafficking networks. And we can do network analysis on the connections between different traffickers. We also have location, a form called a location form. If there's a location, a business, a hotel, a transit hideout or some place that was involved in the trafficking incident, we fill out a location form and then finally we have a potential victim form. And that has to do with the demographic information about the potential victim and her situation, how she was recruited, how he or she was recruited. And it includes a home situational assessment to make sure that we're sending potential victims back only if they have a safe place to go and that they're not likely to be retrafficked. So we now have nearly 90,000 individual potential victim records in our database. We have over 40,000 suspect records and detailed data on nearly 57,000 interception events. And this makes our database, as far as we understand, one of the world's most comprehensive human trafficking database.
And we can use this data to understand trafficking and understand trafficking networks and that leads to actionable insights that can increase our impact.
And so, yeah, that's what I'd say about our data.
[00:04:41] Speaker A: Yeah, maybe. Two other things I would add, John, is you have often talked about how the data that we do have in the anti trafficking space oftentimes comes from rescues potentially, potentially delayed after someone is being rescued. And by the time that that information reaches the light of day, it's outdated, it's old and it's in very small volumes. And I think what's really unique about Love Justice's strategy is that we're on the preventative side and we have volume of information that we are able to analyze at any given time. And I'm curious John, at what point did you realize that data was going to be critical and how did that insight lead to the beginnings of what is now what you mentioned to our knowledge, one of the world's most comprehensive human trafficking data sets. And internally we call it Searchlight.
[00:05:40] Speaker B: Yeah, I don't know if there was a moment and I would also say if anything, it was others besides me that kind of helped me to see it.
Our first intercept record form actually was developed by a staff who I think even proceeds, you might not even know of her, her name was Alex and she created this form and I remember looking at that form and being like, oh, this is good. And I was like, oh, this is needed.
And then I really got, just kind of took ownership of it, was like, oh, we could also ask this, we could also get this information.
And John Hudlow obviously was a early staff who really was big on data and really helped us hone that forum. And then now, yeah, I mean, over the years we've just had this sense of like, we can make this better. We have this, this continual improvement, you know, idea that's sort of built into how we work, where we want to make all of our tools, we want to drive them towards becoming kingdom class.
So every core process, including each of those forums, has stakeholders. And Hannah, you're one of the stakeholders on I think all of those forums. You know, the group of us have been thinking about this for years and thinking about seeing, looking at how that data can be used, what else we need to do, what we're missing. And you know, once a year for the better part of a decade, you know, we meet and we come up with ways to make those forums even better. Not that can make it.
There can be a cost of changing those forms as our data is, our structured data changes.
But it's also even thinking really, really carefully about how, what data we collect and how we collect it over many years and having group of people doing that is kind of, I guess, has been crucial for us to really get that, get the right data and get it into a structure where we can actionize it into impact.
[00:07:37] Speaker A: Yeah, I think like I could be jumping a little bit ahead of myself here, but I think one of the, one of the stories, testimonies, whatever you want to call it, that I find really compelling was specifically talking about the forms and the different iterations of the forms and going from collecting a small amount to maybe over collecting to then zeroing back down of like, no, this is what we're actually using and too much data is not going to be helpful. And so one of the stories that I like to share in terms of how we actually use the data that we have is the red flag waiting project that we had. And I know there's even been different iterations of how we weight the red flags on our IRF forms that informs a strength of case. Can you just share a little bit of how we've use our own data and how we've used data specifically of cases that resulted in a conviction to then inform how we filter through future cases or cases that we're navigating now?
[00:08:40] Speaker B: Yeah, well, it involves data science That I don't understand. So I got to be a little bit careful that I don't say that what I'm saying about that about it is true. But as I understand it, we have used are cases that resulted in arrest as a truth set and then used predictive modeling and as I understand it, machine learning to weight the red flags on our IRF on our intercept record form. And so each of those red flags has a weight that sort of goes towards, you know, the total red flag points is one of the things that monitors use to decide whether to intercept. And ultimately it is not the, it doesn't have the final say, but it is kind of the guidance we give them. And it's part of how we assess the quality of intercepts and flag intercepts that maybe need further follow up and they may need retraining or maybe should not have been classified as an intercept in our verification process.
[00:09:44] Speaker A: Yeah, which I know listeners probably have a varying degree of what they understand about data and how it's used. But what I think is particularly compelling and has always been compelling, compelling for me is in a simple way we are trying to use predictive analysis to help us potentially real time predict in situations that we don't know for sure.
We're doing our best to kind of tease that out, but by the very preventative nature of our work, we're using data and data science to try and really hone in on that. And I think that's really exciting and something that you don't hear I think in our nonprofit anti trafficking industry very often if, if at all. And so I think John, it would be cool just to touch base on what makes our data set unique, particularly in its focus on the preventative data rather than the post rescue intelligence that I referenced earlier.
[00:10:43] Speaker B: Yeah, well, I don't know if there's anything that in and of itself about it being prevented that makes our data unique.
But it's, but it makes it so that the two things that you mentioned earlier, it's timely.
We know about trafficking as it's happening now and we have thousands of intercepts every month. And so the volume and the timeliness is the part that I think is unique about transit monitoring's effect in data.
And I would say the other thing that I think might, that is a little bit unique about Love justice is we just realized that like we need to be very thoughtful, kind of scientific, geeky, you know, like about this, you know, we need to really think well about how. And so we kind of restructured around this identity of Our core competency of searching out the people, places, and tools to maximize mission impact. And so part of that is using data and data science to find those right places to work. And part of that is those tools are often software and data science tools. And so around. When we restructured around that core competency, we kind of turned into something that is not totally distinct from a tech organization where we have a data science department, we have a software department, and our software and data science teams are as, in terms of number of people are about as big as the rest of our program staff combined. So we've kind of restructured the whole shape of the organization around the need to use data really well and to be really thoughtful about how we use data.
[00:12:27] Speaker A: Yeah, that segues really good into my next question, which is that you even referenced this just a second ago. But our global value of be scientific.
How has adopting a scientific mindset not only been important to our work, but shaped the way that we collect, interpret, and act on data in the field?
[00:12:51] Speaker B: Yeah, well, this data, as, you know, Hannah, used to be decide on data, and we realized there's so much more beyond just decide on data, that there is like a scientificness. You know, science has the advantage of.
I mean, just think of how different the world would be if it's. If it weren't for the advantages of science. You know, life expectancies would be about 50 years shorter.
We would not have any of the. We wouldn't be certainly recording a podcast. We wouldn't even have computers or electricity or, you know, so science is uniquely effective at using and gaining knowledge that can, that can lead to tools that are effective.
And it's so effective that, that it's the difference between camping and the modern world. It's the difference between life expectancy of 20 or 30 years old to, you know, 70 or 80. You know, so it is, it is just massively more effective than everything else. And so, but it's.
It is not really effectively used in the field of charity. And so, yeah, I mean, we really, we want to. We want to lead in that. You know, we want to try to be scientific, and that's a value for us. And so in one way, we have been acting on intelligence from our data into impact for many years.
I remember our investigations department would have these network maps, and they would identify this one individual that they believe is at the center of a huge trafficking network, and they would have that map on their wall for months or in some cases even years, and then they would find that individual, and it would turn out to be a big sort of kingpin trafficker. We had a few, we had, I think, two really distinct instances of that in the past.
But even like impact budget ratio is a tool that we use to allocate by impact and to determine which of our stations, which of our even, even which of our monitors, which of our countries have the best impact on the dollar. And then we want to invest more resources in the places that have that the best impact. And so there's ways that we're just beginning to start to turn our data into impact. We've just recently been able to connect a deidentified version of our database to a large language model, which is AI, so we call it Searchlight AI. And we're using that for all kinds of things and among them is creating reports for the various stakeholders. So our tech team gets one, our compliance team gets a report, our anti trafficking leadership team gets one, and each regional steward gets one about there. And it just creates these amazing reports with sort of actionable insights.
So those are just a few. I mean, another one is we're trying to develop a theory of disruptive coverage, which is sort of meant to be a scientific theory about how traffic or how transit monitoring results in increased perception of risk by traffickers and then thereby decreases slavery beyond the number of individuals that we're able to intercept.
[00:16:08] Speaker A: Yeah, I want to dig into kind of what we're currently doing and where you see us going with the integration of AI. But before we do, probably because I've been so in this world for such a long time, I think I also want to highlight just the baby steps that it took us to get here and how part of being scientific is having peer reviews and is doing research. And so you mentioned earlier, Alex was the one who introduced the idea of a form. And then I think of Rachel Wells who is currently helping us redesign how our forms look like. And she went to a library, checked out a book on form design and just spent this deep dive learning about how to design forms in a way that allows for the best collection possible. Cross cultural, like that's such a small example of also being scientific with how we're designing and implementing something. Or there was someone else who came and visited us, I think for, I don't know, a couple weeks, and had experience with auditing and connected with John Hudlow and said, hey, you guys should probably be auditing the data that's getting put into your database. And that was the beginnings of what is now like a pretty robust entry auditing process because we Collect all of our data right now on paper and forms and filling out with pens. I think there's like a world in which we'll move away from that and more towards digital entry, but for, for the time being, it's all manual. And so there was this need of, like, okay, we, we probably need to double check what's being collected is being entered into the database correctly. And, you know, now we have a team of 10 plus volunteer auditors that are helping us audit every country's data every six months. And I, I see it all as a preparation. Like, even though maybe we weren't analyzing the data in ways that you had initially hoped for or intended, maybe, you know, 10 years ago, and you've kind of always said, like, I really just want, you know, 10 actionable insights every month that we can capitalize on. We just never really quite got there. I think AI with the, with the introduction of that and then also building out the tech team, like, we're now able to get there. And I think the reason why we're able to do it now really well is because of all those small and faithful steps that I think line up so well with that value of be scientific that has readied us for this moment that we find ourselves in, that could be really explosive, really energetic. And it's because we invited people to speak into that, into that space and helped us sharpen and develop that over time. And internally, staff taking it on and seeing the value and the importance and just kind of running with it, like, similar to what, what you did and what I did with auditing, what Rachel's doing with form design. So, yeah, I'm such a big advocate, like, when we talk about big ideas like this, like, what does that tangibly and practically look like?
And maybe the other thing, John, that I think would be helpful for your listeners is Searchlight. So Searchlight is something that we created internally. You can maybe talk a little bit about that. Just like, historically, like, where did Searchlight come from? And that's also pretty unique because there weren't any tools that were able to, you know, any tools available at the time for us to do what we were uniquely doing. And so you decided to make it ourselves. And I think that was also a pretty crucial decision at the time. So before we jump into, like, how we're currently using it, I'd love for you to just share a little bit, like, what is Searchlight?
How would you describe it to someone who's listening, who's never seen it, and then how that came to be yeah,
[00:19:54] Speaker B: so Searchlight is our human trafficking database. And it is a tool, a software tool, a database that we.
I guess a database and a software tool that is used for form entry. And so forms are filled out, you know, on the field, and then they'll be submitted electronically and the data entry staff will fill it out and there'll be some verification processes that they'll use as they're entering it into the database that, Hannah, you would be better equipped than me to speak on.
And then Searchlight that we use also to look at our data and see trends in our data. And there's various, we used to call them modules. There's various sort of modules. There's parts of search things that it can do. And so, Hannah, actually, I think you probably better than me, would be able to describe in detail what more of those modules. What would you say are some of the features of Searchlight that are.
Would be most interesting to our listener listeners?
[00:20:55] Speaker A: Yeah, I mean, I think, yeah, it's basically a digital replication of the paper forms that we have in the field that our data entry teams check off or enter in the data through. It's a checkbox or a circle to check if it's present. And then obviously you can export the data to see what is filled out, what is not.
I'm talking about this in very basic plain language. And then some of the, I would say more prominent features right now are being able to see data summary. So is data being entered in a timely fashion?
Is it being collected and sent to our data entry specialists in a timely fashion? What's the lag time?
Do we have original paper forms attached for all of the digital entries? And so it just like very practically at a high level, country by country, even station by station can tell us what gives us a snapshot of entry indicators and collection indicators that really help us identify where the gaps are. So that's one. There's also budgeting, like, I think now we've completely integrated our budgeting request process into Searchlight, which then allows us to automate email reports to the field. And it just has really streamlined, I think, some of our finance processes. So I think I also see Searchlight being both this integration of data collection, data entry, but also potentially HR functions across all of our countries. Keeping track number, keeping track of staff. Have they signed this form, this form? This form, like at a very basic
[00:22:28] Speaker B: level, use the music to disable.
Sorry, Anna, I was just activating. I was just opening up the Searchlight engagement page, which is a new feature where we have these Monitors in our office that tell us how many intercepts happen today. The answer is seven so far in the last 30 days. 17, 1935, all time, 100,181. Although that's actually missing some. That's a tweak. And intercept announcements happen in real time. We can see how many intercepts. It'll go to a scrolling map here. Can I share my screen?
I can.
[00:23:04] Speaker A: That's a good question.
I don't.
[00:23:06] Speaker B: I can.
[00:23:07] Speaker A: I think you can.
But you're also just a reminder, John, our podcast is also audio, so you'd have to talk through what you're also showing and seeing on the screen.
[00:23:19] Speaker B: And I'm just looking at a scrolling map showing in Bangladesh we've had one intercept. In India we've had six today. And I think those are the only intercepts, as the day goes on, that this map will just go between the different places and now it's going to bring up a country spotlight. So that's a new feature of Searchlight.
We're in some ways just beginning to get started with the many things we can do if we. With the data that we've been collecting for many years. So. Anna, I didn't mean to cut you off. I was.
[00:23:48] Speaker A: No, no, no, that's okay. I think it was, I think it was good. I was getting maybe a little too in the weeds, but maybe the one other thing I would say, and this kind of blends really like, it kind of naturally blends into. The next question too is some of the features that we have in Searchlight allow us to track compliance at the highest level of our core processes.
And it's by that that we're able to not to basically not do self reporting but for show where fields are compliant, where they're non compliant, where we need to, you know, build a gap and where we don't. So, John, I know we've kind of, you know, touched on this specifically in the last couple minutes, but I'm just curious if there's anything else that you would have to add on how we're actively using this data to sharpen strategy, improve outcomes and steward donor trust across our programs. Maybe just give like some specific examples of each one.
[00:24:46] Speaker B: And I think there's many ways, I mean, compliance that you just mentioned is key. Like if we, if we didn't track compliance because we have these really detailed sort of operational manuals about how like it's called a transit monitoring center. It's a guidebook for how the transit, transit monitoring station should operate. And there's hundreds of things that you need to do correctly. And our at first, you know, 15 years ago when we first created this, it was like, okay, done, we've got the manual, just give it to the stations. But we realized we need to actually be tracking when. And Hannah, you were the reason that we really realized how crucial that was, that we need to be tracking whether they're doing it at our 50 plus stations around the world, which are the ones that are correctly doing each of those hundred things. And that's become instantly very complex.
And so now we know how to do it in our sleep. We have the NCR score. The average national compliance report score for our countries, I think is 91%. But if it goes down, we know about it, I get a report and we can get into, okay, why, which country's went down, what are they not doing? And I don't even have to look at it. Someone is on it. We have the processes in place through our compliance metrics and tools that if any of those things are not being done well, we're going to know about it and something is going to be automatically enacted to ensure that it improves it. So that's just that compliance is itself one. And if it weren't for compliance, they would not be doing the things that make them effective. And I think we can say that with some confidence that were we not following up on those things and checking them, they would not be happening. And so another tool that is called a case dispatcher. And this is a tool that takes basic case outcomes, things that you want to happen. In order for an arrest to occur, you really need a victim who's willing to testify to a crime. You need to have the bio and location of a suspect, and you need to have police who are willing to file a case. And so for each of those case outcomes, there may be a staff who specializes in that, someone who specializes in going to the police, someone who specializes in talking to victims and trying to convince them to file a case or locating suspects.
And.
But the case dispatcher takes all of our data and spits out prioritized by case outcomes based on factors like how strong is this case, how likely is it to result in arrest, how big is this trafficker and how many of the other steps have been done. And it just prioritizes those case outcomes.
So, I mean, we have a tool that scrapes the web and can fill out suspect forms based on open sources from the media. And we have, you know, we do some social media monitoring. We have a tool that helps prioritize ads that are being used by traffickers to recruit potential victims.
We're developing, we have a heat map of our intercepts where we can heat map the road network to show which transit points have the most trafficking going on. And we're trying to build a predictive heat map that actually can take use open source data to show us where in the entire world are the biggest roads where the most trafficking is happening. And you know the last time we actually had a live version of this it took like more than 24 hours to run on an AWS server to try to generate it.
So donor reports under, you know, identify potential exploitation profiles to be able to better talk about how what types of trafficking are occurring. And all of this, all these tools kind of integrated can and in some cases do interact with one another.
So what we envision is this every day we get more data goes into the tools and potentially every day that's from our intercept, potentially every day we scrape the web and we bring in more data that way and then it can actionize those things out via something like the case dispatcher into actionable, very directly actionable insights that help our team say this is the victim, you need to call first, this is the second, you know. And so our, our vision is that data can greatly and radically increase how impactful we're able to be in our work.
[00:29:02] Speaker A: Yeah, one, one, you said a lot in your response there a lot of like really, really good, good stuff. And one thing that I wanted to come back to and, and just maybe re highlight because I think it just exemplifies this be scientific the uniqueness of our data like it, it embodies all of these things that we've been talking about the last 30 minutes. And that was the heat mapping where you guys took all of the data of the addresses, the routes, the PV numbers and you put it on a map to overlay, hey, what are the most common routes being used currently? And you use that, our teams use that to inform okay, where are we going to pilot next?
And instead of okay, let's just pick a random place that you know looks busy or you know, some people have heard of some things like that was a very strategic, here's where we can pilot in these countries next. Or if we go to a new country that we have some data on, okay, why don't we pilot here? Because we might have some data here like that that is just powerful and strategic and I think just leads to more fruit quick, more quickly than otherwise would be the case and just you know let's figure out like where's the most busy place? You know what I mean?
Yeah, I think any before I, before I move on. Anything you want to add there?
[00:30:23] Speaker B: Well, I'll just say that the ones that we're able to do reliably currently are from our existing data. And there's a bit of a self fulfilling prophecy in the fact that we were set up in these locations and therefore according to our intercepts, these are the places with traffic going on.
And we have used it to, to determine location, we have used these to determine locations. So it is true. But what we haven't been able to do is do a predictive one that looks at open source data for the whole world and says where in all the world should we pilot that? And that is where it will become, if and when we ever are able to develop that, that is when it will become very, very powerful. Because the difference between being in the right place and the wrong place, I mean in some places we intercept more than 500 potential victims in a month.
But it's also typical in a gift place like throughout our history we've had places where you intercept 0 to 1 to 2 in a month.
And so that's at least a 600 times factor in difference of being in the right place and not. So it's really important.
[00:31:31] Speaker A: Yeah, yeah. I think we can't talk about data without also talking about impact bias. So for just a minute talk about how the team intentionally guards against our own biases and how that improves the reliability or hopefully, hopefully improves the reliability and the impact of the data we're collecting.
[00:31:56] Speaker B: Yeah, I mean this is a really important idea that doesn't really, it's not out there in the field of charity to my knowledge. But like people should not trust us to talk about our own impact any more than you trust a drug company to say no, we've done the tests, we've done the trials and it works because we have a bias and we have an irredeemable. And the fact that humans have an irredeemable bias is completely built into the scientific methodology, that same methodology that it's the difference between a life expectancy of 20 to 30 years and 70 or 80 years. And so it is really important.
And so, so we have a verification process. We have sort of recognized this and really because we have a lot of very fair minded scientific minds within the organization.
And it's not necessarily coming from me, it's more coming from those people to me. And my role was to resist it and then recognize that it was right and then grudgingly go along with it and then to eventually just completely champion it because I realized how right it is.
So we have a verification process where two staff verify each intercept, that it meets our definition of when to intercept our criteria. And it does not count or even go into our database until that happens. In fact, those seven intercepts I told you we've had today, those are not counted yet. These are unverified intercepts. So they may go through the verification process and actually not make it into the database.
And we've also done, we're pursuing independent reviews. We've done this in three countries where we have independent people who review a random sample of our intercepts and determine whether they agree with our decision to intercept, whether they believe that these people would have been trafficked but for our intercept. And so that's something we've done in three countries currently. Now we're going to just do it as widely as we can.
In fact, the plan has been to do it in as many countries as we can this year and I think we would like to do it in every country where we work because we can't trust ourselves, we can't trust our own staff. So if we have an independent person that really, it's just a lot different, it's a lot better.
And so, yeah, yeah, I mean this helps us ensure that the data that's in our database is data that we have good reason to believe is about human trafficking and that protects the integrity of our data.
[00:34:19] Speaker A: Yeah, yeah. And the verification process really just scratches the surface of all the different compliance metrics that we have around how to dig into those biases.
And that's something I've always really respected and admired about you, John, is just the willingness to be confronted with that and you intentionally surrounding yourself with people that are not that you're not fear minded, but as the, as the CEO and founder of Love justice, like it is your job to be like excited about this impact and to cast vision and like, yeah, you need to be in that space. And so I have really appreciated how you've surround yourself with people who are doing the verification, the impact bias checks like on the side to allow you to do that, but also do it in a way that when you know, you go out and share about the work of Love justice, it's like, yeah, that, that's, that's true, you know, and, and, or we have a lot of confidence that it is true. You know what I mean?
I think what's probably, I would say
[00:35:19] Speaker B: I'M not fair minded.
And I have an additional. And I have. Not only am I not naturally fair minded, but like compared to some people, but, but also as CEO and founder, I just have an additional layer of bias that is, you know, that, that just makes me incredible. And yeah, that was hard. It was hard for. It's so easy to say now, but it was hard for me to get to the point of just like that is of realizing how completely and crucially how true and how important that fact is.
[00:35:47] Speaker A: Yeah, yeah. And it was probably selfishly motivated by me to be like, I want the thing that I'm like giving, like sacrificing so much for this organization that I'm giving up a lot for, to come and work for, you know, because we're expats, we live in the field, like all these things, support raising. And it was just like if I'm gonna do this, like I wanna do it for something that's really meaningful and something that I know is making a difference. And that's probably been the selfish underlying motivation even of like my skepticism of.
I want to know that it, you know, we can't know beyond a reasonable doubt. But like I want to know that this works and not just because we say it works or it's a self fulfilling prophecy. But John, I think what's. Yeah, what I think is arguably probably the most exciting part of this conversation is how technology is evolving and how we've been integrating AI like with our Searchlight AI to further strengthen our strategy and maximize impact. And what I'll preface by saying before you even answer is that I really don't think we fully, we know yet and we're just scratching the surface. And the little scratches that we're making on the surface are, are amazing and so exciting and like, oh my goodness, this is going to be so cool to see what happens in the next couple of years.
[00:37:11] Speaker B: Yeah, yeah. As you said, we're just scratching the surface. I mean literally right before we got on this call, as I was preparing, I asked Searchlight AI, which is our large language model, to connect to our database, a de identified version of our database, what should we be asking? And I haven't even read the answer because I just thought of the question. But essentially, I mean, asking it for what are the actual insights and also asking it for critical feedback.
And those are the two broad categories. And so we actually used AI to generate a prompt that's designed to detect risk, identify intervention opportunities, recommend actions that increase our impact and operational quality. And, and we do this Monthly we created this really big prompt that we give it that then creates a separate report for each regional steward for their region, for the tech team, for the compliance team.
And different stakeholders are getting the different reports just to understand big picture what's going on in the world of our data that affects you and where are the actual insights, where are the risks, where are the opportunities? And so that is because, again, we're still discretion services. But essentially recently I asked what are the best ideas given all this data for software tools, asking what questions we should be asking.
We haven't used it this way.
We've used it as a trial, just as a test to, say, generate a donor report for this project. And we've not sent that to a donor, but we've at least looked at it and we asked it to analyze trends over time, pick out anomalies. So again, the world, we are just as the world is just kind of beginning to scratch the surface of what AI can do and what it will mean. I think we're in a similar situation where we are just scratching the surface of the question of what it can do for us.
[00:39:11] Speaker A: Yeah, yeah. And even as someone who has been on the receiving ends of some of those reports, I think one day you were kind of navigating or asking it questions about what are some of the most concerning biggest compliance issues that we have.
And the list that it popped out, I think confirmed what we knew to be true, but articulated it at a much bigger level and in a very clear way. So that was also like, okay, this is, it's not, you know, stuff isn't coming necessarily out of left field, but it's positioning it in such a way that allows us to, I think, just really grasp the volume of the compliance issues that we're navigating and making a very kind of clear way forward of, okay, what do we do with, with this information now? In ways that felt very ad hoc for that team, specifically, of like we're, you know, plugging holes that were leaking and, and we had a. We have a really good system for sure, but I, I think just AI is enhancing that system. AI is enhancing the work that we're doing that I think is, yeah, really, really exciting.
John, as we land this conversation, what is one thing you hope listeners take away from this conversation today?
[00:40:26] Speaker B: Yeah, this is a little bit of a hard one, but I would say data is really needed in anything, but it's certainly needed in the fight against human trafficking. We need to understand a thing to be able to do anything about it.
And Love justice has this strategy, transit monitoring, that we believe is the world's only tangible human trafficking prevention strategy. And it's really unique in the amount of impact that it generates in terms of. And the unique, unique cost per intercept is the most amazing part, and the disruptive preventative effect it has on slavery.
And we often almost don't have the time to talk about this really big, impactful side benefit, which is data, which is that data is desperately needed and completely lacking. And so we've quietly built one of the world's most comprehensive human trafficking databases. And knowledge and. Or data is needed in everything for anything to be effective. And so I almost want to say, if transit monitoring just gave us the data and not the intercepts, and it wasn't just a side benefit, it would still, almost still be one of the world's most impactful strategies for fighting human trafficking.
And so I don't know what I hope our listeners do with that.
We live in this world, in this moment, this era, and where everything is being transformed and our world is being transformed by AI. There's a lot of talk about jobs and how is AI going to affect jobs, but just getting a peek at what AI can do with our data is amazing. But it can't do anything unless humans put the data in.
If it weren't for the fact that we set ourselves up to collect all this data for a couple decades or a decade and a half, we wouldn't be able to do this. And so I would say to our listeners, I would just encourage them, dream about the ways that things that you care about most might lack data and how you can change that to make the world a better place.
[00:42:31] Speaker A: Yeah. And I'll tell you what I hope listeners take away from this conversation today, and that's like, I hope you are like, man, love, justice is cooler than I thought they were.
Because I think when people hear about this, John, and you've seen it, we've had people come into the field, they get to hear more of the program type language that you wouldn't necessarily get, you know, in a fundraising packet or report or on social media or even in, in donor conversations, like, this is under the hood. And I feel like whenever people come in who are genuinely curious and they see under the hood, they're like, whoa, this is cool. And this is really powerful. And so I hope that listeners just get a little glimpse of that, of like, man, this is cool. And like you said, that it sparks different ideas in the fight against injustice. You know, that's what we want to do is we want to share these ideas that I think, you know, that make Love justice so uniquely impactful that we're like that. It's like a little bit of the best kept secrets of Love justice and this is one of them. So yeah, one thing I want to mention is if anyone is listening today that has some of the skills set, skill sets that John mentioned of data science, machine learning, like we are always looking for more people to join the team. I think, you know, that's one area where we just have I think a lot of shortage of staff and we're always just kind of looking for the right people to join our team. So I just encourage you guys check out the website and the open positions we have and we also take volunteers. We've had a number of different kind of tech esque volunteer volunteers work with us over the last couple of years as well that have had some pretty good impact. So, so check that out. And John, I really appreciate your time today. Thank you so much for joining us again.
[00:44:25] Speaker B: Thank you Hannah. We are grateful for the generous support of the Love justice community. Please consider joining our family of donors. Learn more at lovejustice NGO.
