The first draft of a police report can now be written by artificial intelligence. For American Public Media, this is Marketplace Tech. I'm Stephanie Hughes. Technology has been transformative for how police officers do their jobs. Think body cameras, speed detection technology, surveillance drones. And now law enforcement departments are trying to figure out if and how they want to add artificial intelligence to the mix. Greg Barber is a science and technology journalist who recently wrote about this for the online publication Proof News. He explained how AI is being used in police work now. Think about things like a lot of analytics. So like you need to go through a whole bunch of surveillance footage or a lot of documents or you need assistance with writing. And so in the case of police officers, this is writing a lot of reports. And that's something that police officers tend to report as being a bit of a pain point in their job, something that takes up a lot of hours. So a lot of these sort of initial products are geared towards reducing the amount of time spent on these kind of tasks that a lot of police officers hold up as being sort of annoying. Yeah. Tell me more about the use of AI in writing reports specifically, which I understand is a huge burden. Tell me how that's meant to work. Yeah, so the primary product that's really kind of AI focused is a product called DraftOne, which is made by a police technology company called Axon. They're really the kind of dominant provider of police technology, including things like body worn cameras. But their first kind of major AI product was DraftOne, which the idea is to take some of the audio from a police incident or an interaction with the community and summarize that into the form of a police report. So does it save police officers time using AI to write the reports? Axon says that it makes report writing go 82% faster based off of reported data from one of its customers. And also it points to a range of other experiences from departments across the country. But there's only been one independent study so far that comes out of a department in Manchester, New Hampshire. We spoke with Lieutenant Matthew Barter, a self-described cop nerd, who likes digging into the data and generating his own data on how police tech works We were shocked to find that it didn save any time And so that leads us to okay why is that number one Number two, could that change over time as we learn these tools more, as our officers become more adept to using AI tools? That could change in the future. But right now, we don't see those time savings. And what was really interesting about the study is that they actually did a follow-up analysis where they analyzed why is it that so many police officers are reporting time savings when actually according to their data, they aren't experiencing them. And they found this big perception gap between officers actually felt like they had saved a lot of time, but then in reality, they hadn't actually saved those minutes. Right. And that has to do with the amount of cognitive load that it sort of got rid of, right? Like it freed up some of the officers' brain space. Is that the idea? Yeah, yeah, which you could sort of frame as a great thing. I think all of us kind of like having freed up brain space, lets us do other things. But at the same time, I'm speaking with Lieutenant Barter. I think one of his concerns was that if you do free up their cognitive load, you're actually kind of taking away from some of the training that police officers are experiencing when they go about writing these reports. You know, this is how you think about the case. You think about the incident. You sort of reflect on the interaction with community members. So he's kind of arguing that that's actually a good thing to experience, potentially. You know, thinking about students using AI in the classroom, for example, it's a similar argument where if you don't go through the kind of basic work of wrestling with the arguments and wrestling with the words, then actually maybe you lose those skills down the line, which could become very important if you need to take that incident into court, for example, and then need to explain why you did what you did out in the field. What did Axon, the company that makes this technology, say in response to your reporting? So they point to the experiences of the officers that they've spoken with. I mean, they work with more than 18,000 law enforcement agencies around the world. Many of them are already using DraftOne. And so they say that we get reports that there are hundreds of thousands of minutes saved. They say that they have the kind of the hard data as well, that these time savings are occurring as expected. We'll be right back. and your financials into one powerful platform. And now they're evolving into QuickBooks Workforce to help you lead your business with confidence and clarity. QuickBooks Workforce combines human intelligence and AI-powered tools, so you get smart automation without ever losing control. Spend less time reconciling and more time deciding what to do next. And as your needs evolve, QuickBooks Workforce evolves with you, bringing together the core HR capabilities businesses expect with the flexibility to adapt your specific needs. Your processes get streamlined and you get precious time and energy back to move forward proactively. Move from reactive to proactive with brand new tools by making the switch to QuickBooks Workforce today. Learn more at quickbooks.com slash workforce. That's quickbooks.com slash workforce. You're listening to Marketplace Tech. I'm Stephanie Hughes. We're back with science and technology journalist, Greg Barber. On AI generally, how big of a business is this? How much are police departments willing to spend? So one aspect of this is that police departments are tying their AI spending into existing technology contracts. As I said, Axon is basically ubiquitous in the US. They already have pretty expensive contracts for other kinds of police tech, whether that's the tasers or the body-worn cameras or the software that they use to manage evidence behind the scenes. So in some senses, this is really sort of just being infused into contracts as they are re-upping, as they're adding the sort of latest technology to their arsenal. But to put a finer point on it, these contracts, I mean, they're typically 10 years or more, can be, you know, 10 to $15 million for a mid-sized department. And that will include typically the body-worn cameras as well as, a whole sort of slate of software that includes AI. What do you think this all means for how police officers ultimately do their jobs and how we, the public, interact with them and their AI tools? Yeah, I think it's easy to look at these tools as being sort of merely about efficiency, merely sort of about reducing pain points in the office, you know, similar to how AI is being framed in workplaces everywhere. But in this case this is tying directly into public safety into potentially use of materials down the line in court A police report for example is a document that will continue to live with a case if it becomes a case or becomes an arrest throughout the life of that case. And so there's concerns from, for example, from public defenders, actually from prosecutors as well, about what happens when a bunch of AI-generated content ends up in the court record. Could that lead to challenges? Could you lead to errors or sort of hallucinations, as they say, that actually could jeopardize a case or it could lead to issues in terms of overcharging someone, for example. I think as well, in sort of the bigger picture, as these tools kind of become part of everything, there's sort of, as part of this broader question about surveillance, about sort of the amount of data that police are gathering and able to analyze, that can be both a great thing in terms of keeping the community safe, But then also it sort of leads to this situation where there's just a tremendous amount of data being collected and analyzed by AI in ways that I think also bring up some concerns from the community as well. So the question moving forward is what sort of policies are going to be in place for the use of this technology? What sort of automation do police officers actually find helpful? So there's a lot of questions moving forward for departments as they kind of take all this on and are sort of beginning to process what these tools actually mean. That's Greg Barber. He wrote about AI influencing for proof news. Daniel Shin produced this episode. I'm Stephanie Hughes and that's Marketplace Tech. This is APM. So many people right now are chasing fast money, not just to feel secure, but to reinvent themselves. And sometimes that reinvention starts to take a life of its own. I'm Rima Hraes, and this week on my podcast, This Is Uncomfortable, I talk with Patrick Radden-Keefe about his new book, London Falling. It's a story of a teenage boy who became obsessed with wealth and what his family uncovered after his untimely death. unbeknownst to them their teenage son had been moving around London with an alter ego he'd had a secret life and he had been pretending that he was not their son but in fact the billionaire son of a Russian oligarchy be sure to listen to This Is Uncomfortable wherever you get your podcasts
