---
id: "workshop:ai-in-police-operations"
type: "InnovateUS workshop"
title: "AI in Police Operations"
url: "https://innovate-us.org/ai-in-police-operations"
date: "2025-10-01"
duration: "60 minutes"
language: "English"
series: "AI for Law Enforcement"
instructors: ["Patrick Poulin","Shane Evangelist","Anita McGahan"]
recording: "https://www.youtube.com/watch?v=waGz9tVFPJ4"
transcript: "automatic, from the recording"
source: "InnovateUS (innovate-us.org)"
licence: "CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/), as stated on innovate-us.org"
---

# AI in Police Operations

InnovateUS workshop · 1 October 2025 · 60 minutes

Join us for this workshop discussion on how AI is changing police operations. Speakers include experts in AI, enterprise software, cybersecurity, and organization design. Participants will learn how newly available software that incorporates AI changes operational decision-making, how AI enables criminality, and the consequences for core functions such as officer deployment, event tracking, policy compliance, and inter-agency collaboration.

## Learning goals

- Learn how newly available software that incorporates AI changes operational decision-making
- Understand how AI enables criminality
- Learn the consequences for core functions such as officer deployment, event tracking, policy compliance, and inter-agency collaboration.

## Speakers

- **Patrick Poulin**, Inspector, Royal Canadian Mounted Police Cybercrime Unit
- **Shane Evangelist**, Chief Executive Officer, NeoGov
- **Anita McGahan**, Senior Research Scientist, The Burnes Center for Social Change

Part of the series [AI for Law Enforcement](https://innovate-us.org/workshop-series/law-enforcement).

Recording: https://www.youtube.com/watch?v=waGz9tVFPJ4

## Transcript

Automatic transcript of the recording: names and terms can be misheard, and the times can be a few seconds off. Each time links to that moment of the video. The transcript alone: https://innovate-us.org/ai-in-police-operations.txt (plain text) and https://innovate-us.org/ai-in-police-operations.json (every segment with its start and end second).

**Anita McGahan**

[0:02](https://youtu.be/waGz9tVFPJ4?t=2) Welcome, everyone. Welcome to the fourth session of AI for Law Enforcement, an Innovate U.S. Workshop series brought to you in partnership with the State of New Jersey, the AI Center of Excellence for Justice, Public Safety and Security, and the Rutgers Miller Center on Policing and Community Resilience. The goal of this series is to equip public safety professionals with the knowledge and tools to use AI, artificial intelligence, responsibly to protect the public, to reduce crime and to promote fairness. My name is Anita McGann. I'm a professor at the University of Toronto and also a senior research scientist at the Burns Center for Social Change. And I'm thrilled to introduce you today to my friend Patrick Poulin, who's inspector at the Royal Canadian Mounted Police Cybercrime Unit and has experience with these issues. You know, our primary focus, Patrick, today is on the ways in which AI is reshaping police operations, both directly and indirectly. I know that you have both direct experience and policy-based experience thinking about cybercrime.

[1:07](https://youtu.be/waGz9tVFPJ4?t=67) You know, AI can enable workflow redesign and, you know, police operations on the ground, but it also demands, you know, thinking about criminality and technical issues about cross-jurisdictional problems and so on. Can you talk a little bit about some of the new demands that, you know, AI and cybercriminality create in your experience?

**Patrick Poulin**

[1:34](https://youtu.be/waGz9tVFPJ4?t=94) Thank you. Yes. Thank you very much, Anita. And thanks for having me today. It's really appreciated to be invited to those events. And yes, because I am part of the cybercrime investigative team in Ottawa here, like we've been faced with quite a bit of a change over time with regards to cybercrime and the use of AI, whether it was from cybercriminals themselves or even for us to look at different options to start looking at the bigger picture and also the bigger amount of data that we kept on seizing, right? So there's a lot of that, you know, like that demand has been changing over time and more so it's always about for us anyways, right? And I can only speak to what I know, but it's changed quite a bit of what we do and how we do it. And for us to start looking at exponential amount of data that has been growing over the last few years with the same amount of people looking at it, it really has increased over time. And so as we know, cybercriminals are using more and more the use of AI when they're deploying attacks on critical infrastructure, in my case, and how we are looking at those cases has been extremely labor intensive, right?

[2:45](https://youtu.be/waGz9tVFPJ4?t=165) So for us, really, this is what has really changed. What it means for us, it's exactly that, is that how do we adapt, how do we convince the people that manage us that we need to look at different tools that are maybe outside of what we are used to and all of that. And so the scaling complexity has increased and the span, because cybercrime is what it is and it's crossing every single border that we know, is adding to the complexities and for us to interpret the data that we are looking at is extremely demanding on a day-to-day basis. So in addition to all of that, the analysis of that data is extremely critical to what we do. And often, if we look at a simple case, we're looking at most likely an analysis of multiple years of looking at the data and AI could help us alleviate those challenges and then doing the same task in a matter of hours, if not minutes, right? So this is what I see as to what the new demands are when it comes to cybercriminality, but also what we do here within the service, right, within the federal policing.

**Anita McGahan**

[3:52](https://youtu.be/waGz9tVFPJ4?t=232) Thank you, Patrick. I see we've been joined by Shane Evangelist, who is the CEO of NeoGov. Welcome, Shane. I want to turn to you in just a second, but before I turn to you, Shane, can I just follow up a little bit, Patrick, with what you said? So of course, you know, AI and the pace of cybercriminality is just creating this incredible demand on your unit and, you know, your unit is not adding people and adding capacity at the same rate as the need for your services is increasing because of the scale of cybercriminality. Can you add to that a little bit of a perspective on how you work with local police services in Canada? You know, cybercrime may show up in a police report in a jurisdiction. You know, how do you connect to local police operations so that you can, you know, open the lid on what might be a much broader problem that's occurring in a particular municipality?

**Patrick Poulin**

[4:57](https://youtu.be/waGz9tVFPJ4?t=297) And when it comes to those municipalities, there really are difficulties for even members to recognize that it is cybercrime, right? So because cybercriminals are getting better at obfuscating themselves and using AI to do just that, often those cases go undetected and it causes a bit of a rupture in the system to say that. And so for us in Canada, we've created the NC3, which is the National Coordination Center for Cybercrime, and where they do the confliction as to what cases exist and maybe cases that are misunderstood by the frontline members. And so through them, through that unit, it eventually gets filtered down. And with the use of AI, in many cases, then they can really collaborate all that information and put it in containers so that we don't duplicate the efforts across the countries, but also across agencies. And as we all know, we're most likely all looking at the same things. And unfortunately, we're duplicating the efforts and the personnel that are looking at the data. So for us, having that central point through the NC3 and looking at the data and maybe categorizing it in better silos, if you put it this way, I think it really alleviates that need for us to look granularly to all of those things that are presented to us, and also relieve the frontline members from that function, because they are really ill-prepared for cybercrime, in my opinion, because it is very difficult cases.

[6:25](https://youtu.be/waGz9tVFPJ4?t=385) And sometimes the evidence exists in places that they've never looked before, right? So having us participate, or at least being incorporated at the onset of the investigation is going to be extremely valuable for those people. And we all know, right, cybercrime is the pointy end of the stick now, whether it's national security, whether it's pure cybercrime, whether it's a domestic assault, everything is most likely started or at some sort of an online kind of a portion of it, right? So cyber is very much what is part of every single file. And so having the NC3 do that, the confliction is going to be extremely relevant to us anyway, so.

**Anita McGahan**

[7:02](https://youtu.be/waGz9tVFPJ4?t=422) For those on the call, please don't hesitate to put your questions for Patrick in the chat. He's emphasizing the challenges of maintaining capacity, both centrally in the Royal Canadian Mounted Police, which performs many functions that would be analogous to federal functions in the United States, and also making sure that members, police officers, sworn officers are aware of what cybercriminality may be happening in their jurisdictions and getting the resources to deal with it. Let me turn now to Shane. Shane, thank you so much for joining us.

**Shane Evangelist**

[7:39](https://youtu.be/waGz9tVFPJ4?t=459) Anita, lovely to be here. I apologize for being a little bit late. We're actually at our user conference. I was just in with a bunch of police chiefs specifically talking about movements in AI and policing.

**Anita McGahan**

[7:52](https://youtu.be/waGz9tVFPJ4?t=472) Amazing. So can I ask you to tell us a little bit about that? What are you hearing about? So just to remind everyone, NeoGov is one of the largest providers of software to police departments in the United States, primarily for managing operations, and especially human resources, deployment of officers, but also a whole range of specialized police functions. Can you talk a little bit about what you're seeing in terms of how police operations are being affected by AI and cybersecurity more generally?

**Shane Evangelist**

[8:32](https://youtu.be/waGz9tVFPJ4?t=512) Yeah. On our front, so for people out there who might know, in the public safety space, we trade under the brand PowerDMS. And the business started as a policy management repository that would make sure people were operating under policy accordingly. One of the first areas we've actually used AI to help officers is in forward deploy scenarios we're testing right now, you might be in route to a child abduction. It's a high sort of scenario with, you probably haven't done this a lot. And so the question is, what are my policy procedures on the way? And you can now, through your body cam, ask for that policy to be right back to you. So what is the policy specifically around engagement of a child abduction when I get on the scene? And I don't want you to summarize it for me because AI can hallucinate, but what I do want you to do is actually read back for me the specific engagement procedures.

[9:34](https://youtu.be/waGz9tVFPJ4?t=574) And so that'll come back to you. So it's very fast from that perspective. And the ability for us to actually use these agents to serve as that sort of intermediary has been, it's going to be great on a go forward basis. The other thing that's quite interesting that we're seeing, again, related to sort of policy and policing is body worn camera footage is sort of, it's a lot, right? And you can't go through all of it. But some of the best training opportunities is in that footage. And truth be told, from a liability perspective, we don't necessarily want to go through all Our policies say when I have an internal investigation or I want to do some training, then I can of it, right? But for the most part, I don't view it all because it's just not practical to do that. pull footage. With AI, it is now very practical to, in fact, examine all body worn camera footage, compare that to behaviors that you want, or compare that to specifically the policy, and then be able to give notes to an officer on that engagement. Now, I would also tell you that we're building it with what we call sort of snap chat feature where those comments and that video will then evaporate.

[10:43](https://youtu.be/waGz9tVFPJ4?t=643) And the reason we're saying that is we've had a lot of feedback from chiefs specifically that says, I just don't want all that hanging out there, but I do want to use it to train. And so this sort of very interesting scenario where you can actually monitor all that video, you can then give quality feedback to an officer, you can then have that officer engage with that, both Axon and Motorola are now in the process of integrating with us to do this work. And so this is really more about how do I make policing safer, but not how do I get the officer? Like, that's the sort of place we want to make sure we straddle. And the way we straddle that is to give you the feedback, but then, lack of a better term, get rid of the evidence that's available if you want to get to it. What's nice about that now is I've got all, I may, I'm not going to engage command staff, but I might engage training. And so I might see what that is. And then I might, A1, we've got this product that allows you to send cards that would say, hey, tell me about a policy and answer a question.

[11:48](https://youtu.be/waGz9tVFPJ4?t=708) If you get that question wrong, we'll direct you to the policy. But in two weeks, we'll send you another variation, all AI driven, of that same question that could be approved by administration. And so what we see from that perspective is sort of retention of that policy go from about 60% to 90%. Again, all of this done without command staff having to engage, I think that's what's impressive about AI, is that we can do a lot of sort of movement of the safety of that officer, both to themselves and the community, by simply using the tools and the footage and the policies that are available.

**Anita McGahan**

[12:19](https://youtu.be/waGz9tVFPJ4?t=739) I mean, you're focusing on improving training and using AI to analyze operational data to try to identify opportunities to support officers, ideally, and to help officers solve the problem Patrick identified, which is even just knowing that what may be happening

**Shane Evangelist**

[12:42](https://youtu.be/waGz9tVFPJ4?t=762) in the field is a systemic problem, or even a problem having to do with cybercrime itself. There's a number of questions in the chat about the technical elements of the tool and service, and we can come to those in a minute.

**Anita McGahan**

[12:57](https://youtu.be/waGz9tVFPJ4?t=777) But I wanted, Shane, can you talk a little bit about how the software or even the analysis that you're doing on different kinds of data that may emerge through body cams, but also through the general process of gathering information, data on police operations. Is there a way that you can more systematically identify cybercriminality occurring in place?

**Shane Evangelist**

[13:29](https://youtu.be/waGz9tVFPJ4?t=809) So most of... I don't have that, Anita. So I wish I had a better tool set. That's not our expertise. I want to be clear. I see a whole bunch of questions in the chat about what is the tool and how does it work. That tool is being developed right now. So that tool is not in market, but it's in process of being put into market. The tool, the question around hallucinations around who bears the liability of that, that's why specifically we're saying hold back the actual text. Don't summarize the text for them. We actually want that text to come directly out of the policy. So those are some of the questions I think that came up in the cyber. My point is specifically around what is AI doing? It's allowing you to digest what was otherwise undigestible. If you want to say what's the theme on AI from that perspective, it's I can digest large quantities of data that was otherwise undigestible because you'd have to have a one-for-one person looking at that and then being able to give feedback on that. The question is how do I safeguard that so we don't get anybody in trouble or we don't And so that is a big focus, I think, of what not only we're doing, but many AI businesses get bad direction.

[14:37](https://youtu.be/waGz9tVFPJ4?t=877) are doing from that perspective. On the cybersecurity side, Anita, I think what Patrick discussed was I think well above my pay grade.

**Anita McGahan**

[14:45](https://youtu.be/waGz9tVFPJ4?t=885) Patrick, did you want to make any observations?

**Patrick Poulin**

[14:49](https://youtu.be/waGz9tVFPJ4?t=889) No, and I think as most of you know, the RCMP is kind of split in two, right? We have the contract policing aspect, which is the frontline uniform members. Unfortunately, it's not the same everywhere, but we have those two categories. of employees and then within the federal policing kind of environment here within the RCMP, having come from the general duty or the frontline policing in my past, I could say how useful that would be to me. And like as a training officer, and I was also like, you know, a training coordinator for the division where I was, whether it's like use of force and all that kind of stuff, those videos would, like in itself, like be extremely valuable to try and make sure that we can train the members better with regards to responding to those calls. So I 100% see it. And I'm wearing two different hats at once here. And I'm sorry, I went on a tangent about federal policing.

**Anita McGahan**

[15:40](https://youtu.be/waGz9tVFPJ4?t=940) Just to, you know, just to make sure everybody on the call is tracking Patrick's, the framework that Patrick is coming from, in Canada, and Patrick, tell me if this is wrong, you know, in Canada, the RCMP provides contract police services to small municipalities, mainly small municipalities that may not have the capacity to support local policing because of their remoteness or because of, you know, their population. And at the same time, the RCMP acts in a way that for example, the FBI in the United States might act. Is that fair? That is correct. Yes. Okay. And Patrick used to be a field sworn officer actually doing police operations, and now is in the cybercrime division. So he's worn both hats. And that's the stance from which you're responding. Can you talk Patrick a little bit about what you're learning about how to experiment with AI to improve what you do?

**Patrick Poulin**

[16:46](https://youtu.be/waGz9tVFPJ4?t=1006) Well, for us, I think the key point is going to be to start slow. I think there's a misunderstanding, people are extremely scared when AI is mentioned, right? Like, it's like when we transition from the typewriters to the computers, a lot of management, we're not necessarily in favour for that change. And I think the same is applied here with regards to AI because this misunderstanding is across the board for me to convince people like well above me as to maybe the need of AI and maybe start looking at things that could induce like you said, Shane, like, is there hallucinations or biases that all those things, those are answers that are difficult sometimes to formulate from a team like mine to say, this is how we're going to manage those things, right? And so and there's also this, like training versus operations versus administrative work, there's all that question that becomes extremely relevant relevant, because are we exposing very sensitive data to the system that the system is hosted somewhere else, and we have no control over versus a system that is our gap or within our control, right? So there's all those kind of levels for us to mitigate and making sure that we address all the needs that management is asking of us. And the risk is obviously something that is of concern. But for me, like when I see that, you know, like where I want to get started is very doing small and pilot projects. And even to try triaging, sorry, it's different to me, it's difficult words.

[18:12](https://youtu.be/waGz9tVFPJ4?t=1092) The triage of those, like extremely extensive amount of data is extremely important, right? Because we are spending years on data that we've seized, and putting people off the floor to just do the analysis of that, right. And so for us, to be able to not only find the evidence that exists within the data set that we have seized, we're saving countless months, I would say, on its analysis. And so for us, having it sourcing to is going to be extremely important, too, because now we could speak and it could be something that's duplicable by human, so that the program itself, when we look at AI here within the cybercrime investigative team, we're looking at tools that are going to be able to generate the sourcing of that information so that a human can speak to where the program went. So when we come, for me to convince my management on the deployment of AI within operations, these are kind of the questions that I'm looking for. But for me, where I'm starting, and this is what we've been doing is doing the R&D if you want to call it that, but it's really about how we're going to deploy it in a safe manner, and on what data set and what controls do we need to put in place to make sure that it's safe, and that it's duplicable by human for when comes court time, and the judge is asking us how did you find that information, and then we can speak to that. And for us, it's extremely relevant that we can do that. So

**Anita McGahan**

[19:32](https://youtu.be/waGz9tVFPJ4?t=1172) that's very resonant with some of the sessions on AI and policing that have preceded this one, you know, that idea that there has to be robustness for the processes in the courts. Can you say anything, Patrick, about sort of scanning for cyber criminality? Does AI help there at all?

**Patrick Poulin**

[19:54](https://youtu.be/waGz9tVFPJ4?t=1194) It will, not within our control, but because then we get into the you know, like almost like a random virtue testing, we're going online and looking for crime. It's not what we do. I know that, you know, like the private industry will most likely do that, whether it's a banking system or those private and like entities, they'll have those things in place, and for which I think will help them report to police a crime that they have perceived based on their analysis, independently as to how that data was created or analyzed, right. So then we can then look at those findings and then source that information, go backwards into the findings of those. But for us, we don't just scan online, unfortunately, and for good reasons, like we don't lack work right now. So we're not looking for more. But when we look at what's coming in with regards to complaints, and like people that, you know, are saying that a crime occurred, then I think there's, at the NC3, a mechanism in place that will be able to filter into what category of crime that is, and who should it belong to based on that, you know, the confliction that will exist within that team.

**Anita McGahan**

[21:01](https://youtu.be/waGz9tVFPJ4?t=1261) Thank you so much. Shane, can I turn to you? Can you, are you seeing any, you know, AI experiments or pilot projects that are exciting?

**Shane Evangelist**

[21:12](https://youtu.be/waGz9tVFPJ4?t=1272) Yeah, outside of cybersecurity, Anita?

**Anita McGahan**

[21:15](https://youtu.be/waGz9tVFPJ4?t=1275) Yeah, and operations very broadly.

**Shane Evangelist**

[21:17](https://youtu.be/waGz9tVFPJ4?t=1277) Yeah, I think, broadly speaking, AI is making a lot of stuff more way more efficient inside policing, generally speaking, like, for instance, we're taking CAD and RMS data, when you do an incident report, and we're using that to help summarize that incident report, and then allowing an officer to go in, and then update that incident report. And so that's sort of super helpful, because I get an officer on the street more than in doing that. There's a, there's a product that we have that's been super helpful around summarizing background investigations associated with an officer, someone who wants to become an officer. And that process, historically, is very complicated to go try to get 20 or 30 applications, and references and bring them in and then summarize those references, and then put a doc together for approval. All that now is super automated. And it's really becoming more of a reviewing process than it is a production process. And I think that's where you're seeing AI make big, big steps, which is this whole concept of before where I had to take multiple areas, bring them all together, and then somehow come up with what my belief my outcome would be. I'm now reviewing that outcome, editing that outcome and saying, I feel very good about this and the speed at which these investigations would take place are sort of speeded up by three, four, five hours at a time. And so that's sort of significant time savings. So there's the efficiency side that we just discussed. There was the side that I just mentioned earlier around being able to review lots of content that we otherwise couldn't review, and then make a decision. If you think about AI, generally speaking, it's basically saying, hey, if I had all this information as a person, I could make this decision, right? That's effectively it. And it's got this unlimited capacity to make that decision. We have an area in our system that we call vitals, which is an early intervention, which I hate saying, because it's had this sort of negative connotation where an officer would be a danger to themselves or the community. And how do you how do you try to intercept that discussion? Because if you were, if you were watching that

[23:18](https://youtu.be/waGz9tVFPJ4?t=1398) officer the entire time, a command staff would be able to say, hey, I'm seeing erratic behavior here. And I should engage and determine how do I how do I how do I reduce that reduce that possibility. We have some software that does the same thing now that takes CAD RMS information, takes sentiment analysis associated when you were deployed. Was there a death? Was there not a death? Was there a death on site before you got there? Did the death happen while you were there? Was it with a child? Was it with a female? Was it with a male? And so all of these different scenarios that happen while you're on the job, as well as you can put things in the system around, are they having financial difficulties to go through divorce if they lose a child or so effectively think about if you have all this data that if any one single person saw all those single points, they could put that together and say, hey, we should really spend some time here. We've got systems that are now being developed that are doing that they're just kind of saying to the command staff, hey, check in, there's a there should be a check in point. I think those are some very positive things that are happening with the use of AI that are effectively taking just large quantities of data that any one given person on one incident could imagine and could digest and say, yes, let's engage. But now you can do it in totality for this for the force.

**Anita McGahan**

[24:32](https://youtu.be/waGz9tVFPJ4?t=1472) I want to give you a chance to ask each other any questions that may have come up that are on your mind. Patrick, did you have anything for Shane?

**Patrick Poulin**

[24:42](https://youtu.be/waGz9tVFPJ4?t=1482) No, I am really interested by all of this. And really, like when you look at the things that we had looked at some of the data like video surveillance, for instance, when we filter that through, like just recognizing by floor, for instance, let's say we have terabytes of information of like video surveillance, AI is allowing us now more than ever to look by floor or when humans are present, for instance, and, and like that type of, let's say a parcel or a purse or like all those things that could be triggered through the use of AI, I think is extremely relevant in this case. And I think this is I think what you've been talking about, like minimizing the data that humans have to look at, based on the needs or maybe judicial authorization that's allowing us to look at certain aspects of the video.

**Shane Evangelist**

[25:26](https://youtu.be/waGz9tVFPJ4?t=1526) Yeah, Patrick's right on. I mean, I'll be very clear on this. The systems specifically on video analysis aren't there yet, because you've got to train that system. So when you engage the weapon or not, there's specific timelines to do that. And being able to match that up is that's complicated. And so you'll see us do that over time, you'll see us train that system over time. And the best place to get the training is you do it, you engage with the officer, the officers know, and I'm gonna explain to you why that's I did engage properly. That's interesting for us, right? Because now I'm going to train the system on that. So effectively, officers will over time train the systems, right. And the nice thing about it, if you could have an officer train that system without penalty, it's even better, right. And so we just have to get our heads around that sort of engagement. And so I love this part, I also think there comes a tremendous amount of responsibility with this part. And we have to be extremely responsible about how we engage and how we build these systems.

**Anita McGahan**

[26:31](https://youtu.be/waGz9tVFPJ4?t=1591) I am, I can't help, especially, I suppose, as someone who's, you know, a business school professor, and he's been dealing with administrative concerns for most of most of my career, there's got to be opportunities to, to, to streamline workflows, right, to just make the administration of policing, you know, simpler, easier, and less onerous. Am I right about that? I you tell me.

**Shane Evangelist**

[26:56](https://youtu.be/waGz9tVFPJ4?t=1616) Yeah, I'll go down one path. If you're in the US, and you want to be an accredited body, whether it's a national accreditation, like clear, or whether it's a state accredited body, there are certain standards for which you have to follow. And those standards get published, and then you have to write your policy against those standards. And then every three or four years, an auditor has to come in and audit those policies to make sure they're in compliance. But then before that audit takes place, an administrator has to go in and audit to make sure they're up to date with the policies and standards that have changed in the last three. And so that's a very cumbersome process. It's just very manually intensive. And the our system is now in beta building where it literally does the audit automatically, it pulls out where the red flags are for both the auditor and the and the person internally, if you want the policy built, it will actually take the deal, it'll, it'll build the policy. And so this administrative function, Anita is, and this is by the way, this is happening across all sort of lines of policing, where it was extremely, is extremely labor intensive, to hit compliance standards, like that's the function, right? The reason it is so intensive is there is compliance standards in place.

[28:13](https://youtu.be/waGz9tVFPJ4?t=1693) And so in order to hit those compliance standards, there has to be a significant amount of review. And so the question is like these agents, I see these agencies, sort of AI agents, effectively can do a lot of this review, what you have to do is now, I'm sorry, they can do a lot of this work to build, you have to now become a reviewer. So the skill sets, interestingly, changing from the skill set of actually building to the skill set of actually reviewing. And so this is this is and quite, and quite interesting, a lot of people don't like doing that. So there's a lot of people saying, but this isn't what I wanted to do. So it's also probably going to be retraining the workforce to want to go down that path.

**Anita McGahan**

[28:49](https://youtu.be/waGz9tVFPJ4?t=1729) Patrick, let me turn to you. Are you, what do you think is coming down the path? Are you excited about, you know, any application of AI or any way that AI can help you do your work more effectively, that you think is going to make a, you know, have a big impact?

**Patrick Poulin**

[29:10](https://youtu.be/waGz9tVFPJ4?t=1750) I think so. I think it is the expectation of a cybercrime unit like mine to be part of that change. And I think this is what's exciting about it for me. Coming from a like a computer scientist kind of background, I really, I think we are the right unit to push this forward within our organization. And of course, it takes a lot, right? Because it's a big change. And so it comes with its many hurdles, as I've discussed before. But I think it is extremely, I think, encouraging that people are finally listening to us. I think this is a big, big thing for us is in effecting change is extremely difficult for any organization. And I think having, you know, been in cyber for some years now, I think it is, we're finally getting the years of people that can affect that change at the higher, much higher level than myself. And that there are, you know, a Looking at maybe its potentials, and we're, you know, like with changes over time with regards to finances and like maybe budgets that have been allocated to the RCMP for federal policing, for instance, it doesn't necessarily grow as fast as crime grows when it comes to federal policing.

[30:15](https://youtu.be/waGz9tVFPJ4?t=1815) And I think this is one of the tools that is going to be extremely efficient, and for its cost, right. And I think that's, we get a lot of people listening to us when we say that there could It's one of those things that we do more with less every day. be savings, right? And I think every police agencies that are connected here today can speak to that is always about doing more with less. And I think this is one of many tools that we should all have on our tool belt. And when it comes to our organization and the RCMP here in Canada, I think it's, it takes a while, but I think we'll get there eventually. And people are listening to the change proposals that we are putting forward at this time.

**Anita McGahan**

[30:51](https://youtu.be/waGz9tVFPJ4?t=1851) Thanks. And I want to turn a little bit to the team here that's online and ask you to just drop your name in the chat if you have a question that you want to ask so I can maintain a queue. Let me begin by asking Sarah to unmute and ask her question. Sarah, are you up for doing that?

* * *

Original page: https://innovate-us.org/ai-in-police-operations · Licence: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/), as stated on innovate-us.org