Shaping the record

Summary

Podcast Summary: Click Here – "Shaping the Record"
Recorded Future News | May 26, 2026

---

Episode Overview

This episode of "Click Here," hosted by Dina Temple-Raston and featuring guest journalists and experts, delves into the growing adoption of artificial intelligence in policing—specifically the use of AI to generate police reports from body camera footage. Rather than focusing on flashy predictions of robotic cops, the episode explores the mundane but critical roles AI fulfills behind the scenes, raising pressing questions about accountability, oversight, and the consequences for justice.

---

Key Discussion Points and Insights

1. AI's Entry into Everyday Policing (00:02–07:42)

• AI Use Cases in Policing:
  • Facial recognition linked to body cams, license plate readers, gunshot detectors, and now, AI-generated police reports.
• The Importance of Police Reports:
  • These reports serve as the "first draft" of history in the justice system, informing charging decisions, court proceedings, and public understanding.
  • Sean Powers: “Once something is in that report, it tends to carry forward right into charging decisions, into courtrooms, and how a case is even understood.” (01:59)
• Promise vs. Reality:
  • Companies like Axon promise efficiency gains—less paperwork, more time for patrol.
  • Initial results from pilot programs (e.g., Edmonton, Manchester) challenge these promises: in some cases, there is zero improvement in officer productivity.

2. Human Memory, Truth, and AI Bias (04:34–07:42)

• The Risks of Over-Reliance on AI:
  • AI-generated drafts may influence or overwrite officers' own memories and interpretations.
  • Concerns about accuracy, loss of human nuance, and what happens if AI "gets it wrong."
• Public Awareness and Oversight:
  • There is limited dialogue around whether communities want these technologies.
  • Sean Powers encourages questioning: “It’s not accusing them of anything…We don’t know where it goes wrong all the time, so we need to ask those questions.” (05:47)
• Potential for More Factual Reports?
  • The dream is greater factual accuracy, but problems remain—transcripts may capture words but miss tone and context.

3. Pilot Programs & Mixed Results (10:11–16:47)

• First-Person Perspectives:
  • Ian Adams (former Utah police officer, now professor) reflects on body cam and AI adoption. He was involved in evaluative pilot programs, offering a pragmatic viewpoint on tech implementation.
  • Matthew Barter (Lieutenant, Manchester Police Department) emphasizes the need for transparent, scientific testing: “I also know it’s extremely important that we test it and we test it kind of in an independent way and in a scientific way." (13:57)
• Axon's "Draft One" Product:
  • Transcribes body camera footage and generates a draft report for officer review.
  • CEO Rick Smith outlines the pitch: “If a police officer spends half of their shift writing reports and we could cut that time in half, that's roughly the equivalent of increasing your force size by 50%.” (14:57)
• Empirical Findings:
  • In Manchester’s six-week trial, no significant time was saved compared to traditional report writing.
  • Ian Adams reflects: “I am surprised…I certainly thought, given my own experience with ChatGPT, we would expect some significant time savings. As a scientist, it's a rare pleasure to be surprised to be totally rejected by the data.” (15:54)

4. Why the Productivity Gains Fell Flat (16:47–18:14)

• Complexity of Police Reports:

  • Most of the work is data entry (witnesses, inventorying evidence, attaching photos)—tasks not easily automated by language models.
  • Not all policing situations are captured by body cam (e.g., phone statements).
  • Officer willingness to use AI tools varies.

• Narrative vs. Data Entry:

  • The actual written narrative is a small part of the labor; existing AI tools don’t streamline the bulk of the tedious, procedural documentation.

5. Potential Benefits: Quality & Consistency (18:14–20:42)

• Support Roles for AI:

  • AI tools can help with proofreading and standardizing reports, reducing omissions or inconsistencies.
  • Helps officers recall details when writing reports hours after incidents.
  • Ian Adams: “Helping officers get all the details right on that first report is something Axon thought about...The advantage of Draft one is it's going to evaluate the entire body camera video and it's going to pull out details that might have been missed.” (19:41)

• Importance of Accuracy:

  • Police reports ripple throughout the justice system—affecting prosecutors, defense attorneys, judges, juries, and the media.
  • "A lot hinges on the police report...as this technology takes hold in American policing, we might expect to see some downstream effects for all those people as well.” – Ian Adams (20:14)

6. Oversight, Profit Motive, and Who Sets the Rules (20:42–25:38)

• Concerns About Corporate Influence:

  • Andrew Guthrie Ferguson raises alarms about for-profit companies, like Axon, shaping the mechanisms of criminal justice.
  • “I worry that the incentives of a company that is...responsible to its shareholders...might run and be in conflict with the interests of a city or a community that might want less policing or less surveillance.” (21:11)

• Lack of Regulation:

  • There are no consistent local, state, or federal regulations guiding use of AI in police reports.
  • Legislative attempts remain patchy and reactive.

• Need for Inclusive Dialogue:

  • Ferguson: “There should be some other stakeholders involved, like, let's ask the community, let's ask the city council, let's ask the people who study AI and know all the problems.” (23:07)
  • Warns against Silicon Valley’s "move fast and break things" mentality, which can disrupt societal norms and privacy.

• Lessons from Body Cams:

  • Adams points out that body cams didn’t produce expected behavioral changes—it’s documentation, not deterrence.
  • “The overall effect of body cameras has probably been to maybe at the margins, very, very small effects for use of force. And that's after a decade of research.” (24:16)

7. The Challenge of 'Datafied' Justice (29:10–36:06)

• Technological ‘Superpowers’:

  • Ferguson describes AI as giving police "a superpower" to aggregate and analyze vast data, leading to unprecedented surveillance capabilities (e.g., tracking all red hats across a city).

• Systemic Risks:

  • AI-generated reports feed into a justice system where 95% of cases resolve in plea bargains—the initial version is rarely interrogated.
  • “If [the police report] has been AI generated, ...it means the entire system is sort of infected with this one document. That’s controlling the whole pathway of a criminal case in ways we haven’t fully comprehended...” – Ferguson (30:55)
  • Errors or omissions in AI reports may never be noticed or corrected.

• Perceived Objectivity:

  • AI-produced reports may appear more authoritative and impartial, inflating users’ confidence and discouraging scrutiny.
  • Ferguson: “The sense that technology gets it right, that it must be an accurate transcription, is something I think we all buy into, even if we know…there are going to be mistranslations and the rest.” (32:46)

• Efficiency vs. Rights:

  • Axon’s vision emphasizes seamless, efficient, interconnected policing tech; Ferguson warns that efficiency isn’t the only value to weigh—privacy, accuracy, and civil liberties matter too.

• Data Control:

  • As individuals, we have almost no say over what becomes evidence in the digital age.

---

Notable Quotes & Memorable Moments

• “If a police officer spends half of their shift writing reports and we could cut that time in half, that's roughly the equivalent of increasing your force size by 50%.”
  — Rick Smith, Axon CEO (14:57)

• “As a scientist, it's a rare pleasure to be surprised to be totally rejected by the data.”
  — Ian Adams (15:54)

• “[AI-generated reports] means the entire system is sort of infected with this one document. That’s controlling the whole pathway of a criminal case in ways we haven’t fully comprehended or thought about.”
  — Andrew Guthrie Ferguson (30:55)

• “The sense that technology gets it right, that it must be an accurate transcription, is something I think we all buy into... I think that's a real danger if you can't have moments in time in the system to interrogate whether it did get correct or not.”
  — Andrew Guthrie Ferguson (32:46)

• “Like, hey, everybody, take a deep breath. We should be really careful about the claims around products before they're evaluated in the context they're meant to be used, not the sort of studio where they were developed.”
  — Ian Adams (25:22)

---

Important Timestamps

• 00:02 – Introduction to AI in policing and report writing
• 01:21 – Types of AI now used in police work
• 03:20 – Introduction to Axon’s Draft One report tool
• 10:11–15:18 – Police officers and researchers on the testing of AI-generated police reports
• 18:14–20:42 – What AI can (and can’t) improve in reports; quality/accuracy issues
• 20:42–25:38 – Corporate influence, regulation gaps, and the need for public dialogue
• 29:10–36:06 – Broader concerns about surveillance, “datafied” justice, and who benefits
• 32:46 – The risk of excessive trust in technological “objectivity”

---

Tone & Style

The tone throughout is investigative, pragmatic, and occasionally wary—emphasizing nuance, skepticism of marketing claims, and the need for real-world evidence. The hosts and guests blend scientific curiosity with lived experience, peppered with humor and accessible analogy (e.g., comparing AI tools to ChatGPT or referencing the mundane reality of police paperwork).

---

Summary Conclusion:
The episode ultimately spotlights the quiet revolution in policing as AI tools move from sci-fi headlines to deskbound realities. While companies and some officers tout the potential for efficiency and improved documentation, field results are mixed—and the stakes for justice are high. Experts urge caution, transparency, and broad stakeholder involvement as society grapples with new forms of technological authority over the “official record.” AI is changing not just how police work gets done, but how truth itself is constructed in the justice system.

Transcript

[Dina Temple-Raston]

From Recorded Future News and prx, this is Click here. Hey, there, it's Dena. A quick note before we start. Twice a month, we team up with WAMU and NPR's One A News Magazine for something we call Cyber Monday. This time, we're looking at how artificial intelligence is making its way into policing. And when one police department put that to the test, it raised a bigger question, because they're using a tool that promises to turn body camera footage into police reports. Faster paperwork, less time behind a desk. But if AI is helping write the official record of what happened, what happens to human memory, who checks it, and what happens if AI gets it wrong? This isn't about futuristic cops. It's something quieter than that, because AI could be creating the paperwork that helps decide who gets charged, who gets believed, and who goes free.

[Sean Powers]

The first voice you'll hear is from

[Dina Temple-Raston]

1A's host, Jen White.

[Jen White]

Dina, welcome back.

[Sean Powers]

Thanks so much.

[Jen White]

Okay, so when we say AI is entering policing, what does that actually mean?

[Sean Powers]

Right now, it has actually slipped into all kinds of aspects of policing. You know, we just finished some reporting on a pilot in Edmonton in Canada that's putting facial recognition on police body cams to match faces to mugshots. There's AI enhanced license plate readers which scan and track vehicles over time. There's gun shop shot detection that uses AI. But the new, new thing that we were focusing on was the AI generated police reports. And as you said, it may sound

[Dina Temple-Raston]

small, but it's actually one of the most important documents in the entire justice system.

[Sean Powers]

They're a first draft of what happened. And once something is in that report, it tends to carry forward right into charging decisions, into courtrooms, and how a case is even understood.

[Jen White]

But as you said, this version of what happened can have legal ramifications. So talk about the significance of that.

[Sean Powers]

Well, let's say that an officer doesn't remember all the details of something that happened or remembers it in a certain way.

[Dina Temple-Raston]

They're human.

[Sean Powers]

So they would normally write out that report, and then they would check it and recheck it. What this first draft does is kind of the same way that ChatGPT, if you're using it, it's really easy to lean on it too much. The concern is that the officers will lean on this too much. That draf one's version of events colors the way they remember the events. And because of that, it wouldn't be exactly accurate.

[Jen White]

So for the companies that are developing this technology, what are they saying this technology can do? What are they promising?

[Sean Powers]

It's the big promise of AI Right? The pitch is efficiency. There's a company called Axon. They make Tasers and body cameras. And they've developed a tool that takes body camera footage, transcribes it, and then turns it into a draft police report. So instead of an officer sort of sitting down after a shift and writing everything out, this system does what they

[Dina Temple-Raston]

call a first pass or a first

[Sean Powers]

draft, and then the officer edits it. And the promise is less time typing, more time policing. But it turns out that officers do not love AI as much as we think they might love AI and that

[Dina Temple-Raston]

there are a lot of little details

[Sean Powers]

that go into a police report. So it's not just what transpired when you pulled someone over or arrested someone. There are all kinds of exhibits that you have to back and number and put that into the report. So what they found is they expected Axon said they thought there would be a 40% increase in productivity. And what they found at some of these pilots that they used is that

[Dina Temple-Raston]

there was actually no improvement.

[Jen White]

Dina, whenever you bring these stories to us, I'm always like, how did you start looking at this story in the first place? What brought it on your radar?

[Sean Powers]

We had heard that there were these different pilots with technology, and clearly, Click here. We're all about how technology is changing everything about our lives. And we'd heard about the Axon pilot. We track Axon pretty carefully. Axon is trying very hard to find a way to marry technology and policing and to make it more efficient and to make it more fair. And it's just sometimes they're a little bit out there with wanting to use the technology.

[Jen White]

You were a counterterrorism reporter at NPR for many years. As you're looking at the way technology is being integrated into policing, are there any similarities with what you found when you were on the counterterrorism beat?

[Sean Powers]

A huge number. And in fact, the reason why I ended up in tech was because that's where counterterrorism was going. It started with this sort of taking all this information and hoping you can see a pattern so that you can sort of. This sounds very sci fi, but that you could figure out who might have a propensity to be a terrorist. And so the concern is that, you

[Dina Temple-Raston]

know, that's a test.

[Sean Powers]

But what's going on with policing and technology is very much touching all of

[Dina Temple-Raston]

us in big ways. You know, license plate readers, it sounds

[Sean Powers]

like it's pretty benign, but in fact, knowing what your license plate is and tracking it across state lines and somewhere Else gives them all kinds of information

[Dina Temple-Raston]

that you don't really even think about.

[Jen White]

Well, who gets to decide whether a technology is used in policing here in

[Sean Powers]

the US that's one of the problems is. And that's really the tension that we're trying to explore in our story. Is this stuff that just gets applied

[Dina Temple-Raston]

and you get no voice in it,

[Sean Powers]

or is this something where you can stand up and say, no, I think this is a little bit too intrusive. I think this is more than what I want in my community and there's not enough of that dialogue. So one of the things we're trying to do is raise awareness so that you do ask, hey, are you guys using facial recognition with body cam cameras in some way? Are you using AI to write police reports? What are you seeing when you do that? And it's not accusing them of anything. It's. This is very new technology. We don't know where it goes wrong all the time. So we need to ask those questions.

[Jen White]

We got this question from HD who says if AI interprets police body camera footage, could we get more factual reports?

[Sean Powers]

That's the dream, right? Is that there would be more factual reports. But again, think about if you just read the transcript of something. You know, as an audio person, reading a transcript of something versus listening to it is completely different. And tone of voice is really important. But, you know, think about how the debate has changed right now that every police are wearing body cameras when they aren't on.

[Dina Temple-Raston]

When they say, hey, sorry, I forgot to turn it on, people go, hmm, that seems weird to me.

[Jen White]

Well, that brings us to the reporting we're going to hear from you. You talked to some of the police departments testing this new technology. Just give us a little bit of a preview. Are they embracing this or are you hearing a more cautious approach?

[Dina Temple-Raston]

Both.

[Sean Powers]

So, interestingly, we found our own pilot program that Exxon didn't provide for us, and they were pretty skeptical about how much it had really changed and how it had really saved any time at all. In fact, they found that it was Manchester. It didn't save time in the least. We talked to another person in Florida

[Dina Temple-Raston]

and he said his cops loved it.

[Sean Powers]

So I think the jury's still out.

[Jen White]

Coming up, Dina and her team look at AI generated police reports. Who's using them, how do they work?

[Ian Adams]

And.

[Jen White]

And are they saving time or quietly reshaping the record? That's just ahead.

[Dina Temple-Raston]

Support for Qlik here comes from Servil. Every company says AI will make employees more productive, but most employees are still stuck waiting on it, waiting for app access and password resets, waiting for someone to fix a laptop issue so they can get back to work. That operational drag adds up fast, and IT teams are overwhelmed trying to keep up. Servl was built to automate that work. You describe what you want in plain English and Servl builds it for you. No complicated workflow, no consultants, just faster support and fewer tickets, slowing everyone down. The platform is designed to eliminate repetitive tickets so it can focus on strategic work instead of constant firefighting. The company guarantees customers can automate 50% of it tickets. Learn more or start a free four week pilot at cerval.com clickhere that's S-E-R-V-A L.com clickhere servil.com clickhere Support for Click Here comes from NPR's Planet Money podcast. Planet Money has a knack for taking big, complicated stories and making them feel human. Take the conflict between the US And Iran in a recent episode. The show followed a Seattle comic book publisher trying to track down two comic books stuck on a container ship in the Strait of Hormuz. To explain what was happening, Planet Money called a man in Tehran and he described Iran's strategy like 5 seconds to pay or your ship doesn't get to pass through. Just like that, a geopolitical standoff turned into something that you could actually picture. That's the trick Planet Money pulls off over and over again. It funds the people living inside these enormous economic stories and through them helps explain how the world really works. From global shipping to sanctions to why

[Sean Powers]

Pokemon cards are suddenly worth so much

[Dina Temple-Raston]

money, Planet Money makes complicated things feel surprisingly clear. Follow NPR's Planet Money podcast and understand how money shapes the world.

[Ian Adams]

There's technology in the abstract, right? Its promises, how people think it might be used. But when it meets the realities of policing, it almost invariably becomes changed. It's almost a never ending puzzle of complications and difficulties in implementation. And it's always going to be interesting.

[Dina Temple-Raston]

Ian Adams spent more than a decade as a police officer in Utah, which was long enough to watch a whole raft of police technologies cycle through the department. Some stuck, some didn't. But every now and then something would show up that changed the job. And for Ian, one of those moments was 2011 when he strapped on a body camera. Back then, most departments had not even heard of them yet.

[Ian Adams]

They were testing units. So by the time they became a big part of the national conversation, I was pretty experienced with them. I sort of heard people, especially academics, talking about the camera in a way that didn't make a lot of sense to me.

[Andrew Guthrie Ferguson]

Keep your window down. Keep your window down. I'm gonna get you out of the car.

[Dina Temple-Raston]

But for Ian Adams, that just didn't ring true. He didn't think the cameras would end up helping that way.

[Jen White]

Get out.

[Dina Temple-Raston]

When we tell you to do something, you do it.

[Sean Powers]

This is the body cam audio from

[Dina Temple-Raston]

the arrest of Tariq Hill in 2024. He used to be a star wide receiver for the Miami Dolphins. And he was pulled over just outside the stadium, apparently for reckless driving. And as you can hear, the presence of a camera didn't exactly soften the

[Andrew Guthrie Ferguson]

moment you understand not what you want,

[Matthew Barter]

but what we tell you.

[Andrew Guthrie Ferguson]

You're not confused by.

[Dina Temple-Raston]

The 2010s, body cameras had become the defining technology of the police reform movement. But Ian never thought they would actually change behavior. He thought the cameras would just document it.

[Ian Adams]

I saw it as a tool that helped me quite a bit. You know, it helped protect me in the cases of spurious complaints, and it helped me capture what I was seeing out there in the world as a police officer, which is very, very interesting.

[Andrew Guthrie Ferguson]

Right.

[Dina Temple-Raston]

Ian is now an assistant professor at the University of South Carolina. And when the hype around AI and large language models began, figured it was only a matter of time before it showed up in policing too.

[Ian Adams]

So I knew right away there's. This is a really interesting example of technology being adapted right into the workplace, the police workplace.

[Dina Temple-Raston]

Because if cameras could capture what happened, the next step was obvious. What if they could write it down? Enter Draft one, a product made by the police technology company Axon. And what it does is pretty simple. It takes body camera footage, transcribes it, and then feeds it into an AI system that generates a first draft of a police report one officers can edit and file. The idea is straightforward. Save time, cut paperwork. Pretty soon after Draft one hit the market, Ian got a call from the Manchester Police Department in New Hampshire. They wanted him to do a simple thing. Help them see if Draft one lived up to its promise to cut down on the time it takes to write police reports.

[Matthew Barter]

I am heavily involved in evidence based policing, and there's a lot of nerds like myself, as we call ourselves the Nerd herd.

[Dina Temple-Raston]

Officer Matthew Barter is a lieutenant at the Manchester Police Department. Manchester has about 250 officers and it's considered a mid sized department. And according to Matt, while it's got its fair share of both suburban and urban problems, it's also big enough to provide a statistically significant study.

[Matthew Barter]

So as much as I'M excited about a lot of this technology. I also know it's extremely important that we test it and we test it kind of in an independent way and in a scientific way.

[Dina Temple-Raston]

To hear Axon tell it, this was all about getting police out of the station and onto the streets, interacting with the community.

[Rick Smith]

About half of a police officer's day is spent doing data entry bureaucratic tasks.

[Dina Temple-Raston]

This is Rick Smith, the CEO of Axon.

[Rick Smith]

Every time a police officer stops somebody, they're filling out a new report and all of these things add up over time and it's beginning to squeeze out the actual job that these people are out to do.

[Dina Temple-Raston]

Draft one is supposed to help solve that problem by transcribing the audio captured in a police body cam. It doesn't capture tone of voice. It isn't great if there's lots of crosstalk, but the idea is that it can produce a first draft of a report and then police only need to fill in the blanks. Axon Smith again.

[Rick Smith]

Boy, if we could just help extract the police report from all the things an officer says and hears during the interaction, that would be tremendously valuable. If a police officer spends half of their shift writing reports and we could cut that time in half, that's roughly the equivalent of increasing your force size by 50%.

[Dina Temple-Raston]

So that's the dream, or at least the Axon dream. So Ian and Matt from the Manchester Police Department decided to run a simple control test.

[Ian Adams]

We ran this experiment where half the officers were able to use the draft one product, half had to use just traditional report writing and they let the

[Dina Temple-Raston]

experiment run for six weeks.

[Ian Adams]

And at the end of that, when we compare the duration of how long it took them from start to finish, from the time they opened a report to the time they submitted it, there was no significant effect on time savings.

[Sean Powers]

And were you surprised by that?

[Ian Adams]

I am surprised. I certainly thought, given my own experience with ChatGPT, we would expect some significant time savings. As a scientist, it's a rare pleasure to be surprised to be totally rejected by the data.

[Dina Temple-Raston]

Now granted, this is a single six week trial in just one department, so more research needs to be done. But at least in this one early study, the idea that AI generated reports will free up cops to do their jobs well, it didn't quite work out that way, which Matt in Manchester says was unfortunate, but not totally surprising.

[Matthew Barter]

I never, I guess, really overly get my hopes up on these things. I've also seen, you know, enough studies that things aren't always working out the way people believe they're going to.

[Dina Temple-Raston]

Ian and Matt spent a lot of

[Sean Powers]

time trying to figure out why their

[Dina Temple-Raston]

Manchester experiment turned out the way it did. And among other things, it has to do with the police report itself. A lot more goes into it than the who, what, where of a call. The bulk of the report is actually data entry, lots and lots of data entry. For example, let's say an officer on the scene takes four witness statements, collects 15 pieces of evidence, takes a bunch of photographs and then goes back to the station.

[Ian Adams]

Every person he or she talk to is a piece of information that needs to be added to that database. Name, date of birth, social address, phone number.

[Dina Temple-Raston]

The witness statement has to be scanned and then attached to the report as a PDF.

[Ian Adams]

Every piece of property that was held needs to be entered into evidence physically. So they're going to have to go seal it in an evidentiary bag and

[Dina Temple-Raston]

that gets a number and then has to go into the report as well. Another issue is that not every interaction between officers and the public has a body cam video that AI can use to generate a report for an officer who's on desk duty and is taking phone calls. Well, there's no body cam footage for that. There is also, of course, the possibility that not every officer wants to use AI. Put that all together and you might not get the same time savings that something like Draft one says it could provide.

[Sean Powers]

Do you think there was sort of this maybe, lack of understanding that maybe the bulk of the time wasn't writing the narrative, but is in fact all these other things?

[Ian Adams]

Oh, I think that absolutely has to be part of it.

[Dina Temple-Raston]

When we asked Axon about Ian's study, they were quick to say that this was just one department's experience. And they told us they've heard from plenty of officers who say Draft one is saving them tons of time. Case in point, they put us in touch with the technology sergeant at the Fort Collins Police Department in Colorado, a guy named Bob Younger. And Bob said his department loved it. But even if it can help some departments, Ian thinks generative AI is better suited to lower stakes, nuts and bolts work. Things like proofreading and tightening language and improving consistency across different reports.

[Ian Adams]

You know, officers aren't generally hired for their immaculate writing skills. If you go into any agency and talk to the sergeants, they're going to tell you some of their frustrations about police reports as they're having to revise them, they're having to add to them.

[Dina Temple-Raston]

And helping officers get all the details right on that first report is something Axon thought about. Noah Spitzer Williams runs the Draft one team, and he says that humans forget things all the time, even when they aren't under stress. And Draft One software program helps with that.

[Matthew Barter]

It reminds them of things that they might have forgotten because many officers, they might not actually write that report until several hours later. And now it's two in the morning. They just want to get home. But they got three reports they got to get through before they can do that. And, you know, we're all human, right? They just might not remember all the details. But the advantage of Draft one is it's going to evaluate the entire body camera video and it's going to pull out details that might have been missed.

[Dina Temple-Raston]

Ian says that kind of improvement would not only help the police, but also everyone who relies on the quality and accuracy of their reports.

[Ian Adams]

A lot hinges on the police report from prosecutor decisions, how defense attorneys approach this, how defendants themselves think about their role in the case against them. Juries, judges, even the media and academics consume police reports to one degree or another. And so as this technology takes hold in American policing, we might expect to see some downstream effects for all those people as well.

[Dina Temple-Raston]

All the more reason to make sure AI generated police reports are accurate and fair. Beyond just whether or not products like Draft one save time. People worry about for profit companies getting so involved with policing that they actually dictate how it's done. Just ask Andrew Guthrie Ferguson. He's a law professor at American University and at one time was a public defender.

[Rick Smith]

I worry that the incentives of a company that is, you know, legally fiduciary, responsible to its shareholders to sell and grow and get bigger and bigger might run and be in conflict with the interests of a city or a community that might want less policing or less surveillance or fewer, you know, bells and whistles on the police body cameras.

[Dina Temple-Raston]

The maker of Draft One, Axon, is a publicly traded company, which means they're practically required to maximize shareholder value. In other words, make as much money as possible. Axon CEO Rick Smith told us that the company aspires to be in every part of the judicial process to do for policing what Apple did for smartphones. Andrew, for his part, says we should think carefully about what happens when one company wields that much influence over such an important part of society, particularly if there aren't the regulations in place to govern and hold it accountable. When it comes to setting rules for new tech generally, Congress has a pretty mixed record. It waited to set up the guardrails for the Internet and ended up providing immunity to tech companies and lawmakers are still trying to free themselves from that. They're trying to do better with AI by holding hearings and trying to regulate its use in some narrow instances, like in political ads. But no regulations exist for state or local police. And that worries Andrew. He's written law review articles trying to raise the alarm.

[Rick Smith]

There's no place to ask these hard questions. It shouldn't be law professors that are writing law review articles that no one reads about, like, what are the hard issues here? Like, there should be a process. Wait a minute. You want to change policing by introducing an AI that says the police report?

[Dina Temple-Raston]

He says this shouldn't be a conversation between a for profit company and a police department.

[Rick Smith]

There should be some other stakeholders involved, like, let's ask the community, let's ask the city council, let's ask the people who study AI and know all the problems.

[Dina Temple-Raston]

He worries about how Silicon Valley's move fast and break things approach might apply to policing.

[Rick Smith]

The things that we're breaking are some of our structures of society and our privacy rights and our ability to be secure and not be under constant surveillance. And I just don't think we've had the conversation we should be having about those technologies.

[Dina Temple-Raston]

Ian Adams agrees. And in the meantime, he cautions police departments to take it slowly and thoughtfully. That's in part because of what he discovered in Manchester, but also because he's been here before. Remember, he was one of the officers involved in the early adoption of body cams.

[Ian Adams]

I think even the policy and public view of body worn cameras was that they were tools of deterrence, that they were going to deter officers from using force or deter officers from doing something bad.

[Dina Temple-Raston]

Ian has always been skeptical that police behavior would change, and subsequent studies proved him right.

[Ian Adams]

When you look across 80 or 100 of these types of experiments, then it allows us to get a really good grasp of the overall effect. And the overall effect of body cameras has probably been to maybe at the margins, very, very small effects for use of force. And that's after a decade of research.

[Dina Temple-Raston]

It turns out body worn cameras really just provide documentation after the fact. So it provides the possibility of accountability and transparency if police departments decide to avail themselves of it. So the net effect of AI on police reports or police work more generally may be something we never expected. Yet it is being quickly adopted before we have time to figure that out.

[Ian Adams]

I just feel like we're such at the beginning of this next second phase of sort of body camera and AI, but they're already being widely adopted across the US like we're always behind that adoption curve.

[Dina Temple-Raston]

Technology is great, ian says, but we tend to embrace it before it's been tested and married to regulation and policy.

[Ian Adams]

Like, hey, everybody, take a deep breath. We should be really careful about the claims around products before they're evaluated in the context they're meant to be used, not the sort of studio where they were developed.

[Dina Temple-Raston]

His point is this it takes a long time to sort out exactly how new technologies will affect the real world, much less the complex environment of American policing.

[Jen White]

Up next, who controls the data when big tech enters the world of policing? I'm Jen White. We'll be right back.

[Dina Temple-Raston]

Support for Click Here comes from Quince Lately, I've been more intentional about what I wear day to day, leaning into pieces that feel effortless, comfortable but but still put together. It just makes getting dressed less of a chore. And for a while now, Quince has been my go to. The fabrics feel elevated, the fits are flattering, and everything just works without overthinking it. Quince makes it easy to refresh your everyday this spring with pieces that feel as good as they look. They use premium materials like 100% European linen, organic cotton and ultra soft denim. Their lightweight linen pants, dresses and tops start at $30 and are effortless, breathable and easy to wear on Repeat. Everything at Quint's is priced 50 to 80% less than similar brands. They work directly with ethical factories and cut out the middlemen, so you're paying for quality and craftsmanship, not brand markup. I just got a Quince bathing suit that looks like one of those expensive European brands but for a fraction of the price. Refresh your everyday with luxury you'll actually use. Head to Quince.com clickhere for free shipping on your order and 365 day returns. Now available in Canada too. That's Q-U-I-N-C-E.com clickhere for free shipping and 365 day returns. Quince.com click here support comes from wise

[Recorded Future Representative]

the smart way to manage the currencies you need around the globe. Fed up with losing out to hidden fees when you send your money abroad with your everyday bank? Choose the smart way Wise you can count on the exchange rate you'd usually find on Google. No unwelcome surprises. Plus, ditch that where's my money feeling. Most transfers arrive in under 20 seconds. Join millions, saving billions on hidden fees. Be smart, get wise. Download the wise app today. Ts and Cs apply.

[Jen White]

If there was a big red button

[Andrew Guthrie Ferguson]

that would just demolish The Internet.

[Jen White]

I would smash that button with my forehead.

[Matthew Barter]

From the BBC, this is the Interface, the show that explores how tech is rewiring your week and your world. This isn't about quarterly earnings or about tech reviews.

[Jen White]

It's about what technology is actually doing

[Matthew Barter]

to your work, your politics, your everyday

[Dina Temple-Raston]

life, and all the bizarre ways people are using the Internet.

[Matthew Barter]

Listen on BBC.com or wherever you get your podcasts.

[Dina Temple-Raston]

Hey there, it's Dena again. The scary thing about AI and policing isn't really the robots. It's the paperwork, the databases, the quiet systems that decide what counts as truth. And once those systems start talking to each other, policing begins to change in ways that are harder to see. You're listening to one of our regular Cyber Mondays with WAMU's Jen White. And we'll get back to the show. The next voice you'll hear is Jens.

[Jen White]

Now back to today's installment of our Cyber Monday series. Joining us now to talk about how new technologies are transforming policing and the criminal justice system is Andrew Guthrie Ferguson, and he's a law professor at the George Washington School of Law and author of the new book, you, Data Will Be Used Against Policing in the Age of Self Surveillance. Andrew, welcome to the program.

[Andrew Guthrie Ferguson]

Thank you so much.

[Jen White]

Now, we just heard this story from Dina Temple Rastin focusing on using AI to write police reports. And that's just one of a host of technologies you touch on in your new book. You say this new tech is transforming policing and even the criminal justice system more broadly. So what happens when policing becomes increasingly datafied, more automated, and mediated through these systems people don't fully understand?

[Andrew Guthrie Ferguson]

I think the one thing that we've seen is AI is giving police a superpower. It is changing the scale, the speed, the aggregation of being able to use information in different ways. Like take a video camera. We have video cameras on the street all of the time. We've gotten used to them. But now that video camera has video analytics running over it. So every object on that video stream can be identified. Man, woman, child, car, bus, tricycle. And the way we can do it is AI. Right. We've trained the AI systems to be able to identify those particular objects as something. Now expand that one video camera to hundreds and thousands, all feeding back to a real time crime center where you can now track through a city all of the tricycles or bicycles or vans or red hats.

[Jen White]

It seems like this technology has the potential to either or both expose problems in our system or create new problems. What are you seeing right now?

[Andrew Guthrie Ferguson]

One of the real Concerns with AI police reports is it is the document that police have to rely on. Obviously, prosecutors have to rely on. Judges have to rely on a preventive detention system where they're given the affidavit. And in a case of our current criminal justice system, where 95% of cases result in plea bargains, the police report almost never gets interrogated. So here you have this, like, document that we all rely on. I was a public defender before teaching, everyone relied on the police department police report. But now, if that has been AI generated or partially AI generated, it means the entire system is sort of like, infected with this one document. That's controlling the whole pathway of a criminal case in ways we haven't fully comprehended or thought about. Well, what does it mean? And that might not be the fault of the technology, might actually be the fault of our reliance on a police report to do so much work.

[Jen White]

You used the word infected. That's a strong word.

[Andrew Guthrie Ferguson]

I think it's interesting because if there's an error there, you may never get to see whether it can be cured later on. Right. There may be a no moment in time where we go back to check the videotape to see if we got it right. It may all be based on whether there was an omission, an error, some kind of concern, mistranslation in this initial document, that could change, like whether someone gets, like, a diversionary program, whether they get to get a good plea bargain, whether they get released before trial. And it could all be because of some issue in that police report.

[Jen White]

One thing that struck me is that an AI generated report, it may sound polished, it may sound authoritative, it may even sound objective. So what concern, if any, do you have that the end users of these reports, the justice, the judges, the prosecutors, in perhaps rare cases, juries may trust these documents more because of a presumed sense of impartiality?

[Andrew Guthrie Ferguson]

I think that's fair. I mean, the sense that technology gets it right, that it must be an accurate transcription, is something I think we all buy into, even if we know in the back of our minds, of course there are going to be mistranslations and misunderstandings and the rest. But I think we have this general deference to technology, especially if we don't feel like we know better. And I think that's a real fear as it goes through that, well, this must have gotten it correct because the technology did it, the AI did it. And I think that that's a real danger if you can't have moments in time in the system to interrogate Whether it did get correct or not.

[Jen White]

Well, Axon is one of the major companies developing this broad spectrum of public safety technology. I want us to listen to part of a promotional video that describes their vision of a future that incorporates license plate readers, surveillance cameras, drones. Basically a system that's completely connected and seamlessly shares data in real time.

[Andrew Guthrie Ferguson]

91 1, this is Michael Harris with Secure Solutions Global Security Operations Center.

[Dina Temple-Raston]

We are currently witnessing a crime in

[Andrew Guthrie Ferguson]

progress at our store location at 1201 North Wilmington.

[Ian Adams]

The crime appears to be a robbery in progress.

[Sean Powers]

This is a stolen vehicle.

[Rick Smith]

Let's get a drone out there.

[Jen White]

So the video really focuses on the seamless, efficient criminal justice system, from identifying a crime to writing a report. But from your perspective, how much should efficiency drive the design of the criminal justice system?

[Andrew Guthrie Ferguson]

Well, I think efficiency is the driver because the police are asking Axon and other companies for that efficiency. I think citizens have other questions they might want to ask about accuracy and completeness and being able to protect privacy and civil liberties. As that drone flies over to the scene, is it recording? What are the rules about recording? Who can get access to it? What happens if there's an incident of police brutality or any of those things? But I think that efficiency has always been the push. And watching Axon sort of AI policing future is. It almost feels like a movie. It is objectively impressive in the sense of what you can do, what you can see in all of these real time crime center technologies all being linked together in a really, really powerful centralized command center. Are there places the drones won't fly? Are there places that cameras shouldn't be on? Do you really want it outside a medical clinic or a church? Who gets access to this? Why isn't city data, as opposed to the why would the police get to control who has access and who has not? And those are open questions that I think cities should be interrogating about. If we are investing in these large scale new surveillance centers, what are the rule sets before they get turned on the people?

[Jen White]

So very briefly, how much control do we have as individuals right now to control what becomes evidence and what doesn't?

[Andrew Guthrie Ferguson]

Almost none. Honestly, almost none. I mean, you can choose what to put in your home and put on your door, but generally speaking, with a warrant, almost everything you create in a digital age can be available to the police with a warrant and many times without a warrant.

[Jen White]

That's Andrew Guthrie Ferguson. He's a professor of law at the George Washington University Law School and author of the new book, your data will be used against policing in The Age of Self Surveillance. Thanks so much.

[Andrew Guthrie Ferguson]

Thank you.

[Jen White]

Today's producer was Alison Brody with help from Click Here's Megan Dietre. And this program comes to you from wamu, part of American University in Washington, distributed by npr. I'm Jen White. This is one A that was part

[Dina Temple-Raston]

of our conversation with one A host Jen White about how AI is changing how police operate across the country. You can hear the full [e-mail]

[Sean Powers]

click

[Dina Temple-Raston]

here is a production of recorded future news and PRX.

[Sean Powers]

Today's show was written and produced by

[Dina Temple-Raston]

Megan Dietre, Sean Powers, Erica Gaeda, Zach Hirsch and Casey Georgie. It was edited by Karen Duffin and Sarah Covedo and Fact checked by Darren Ancrum. Original music is by Ben Levingston with additional music from Blue Dot Sessions. Our staff writer is Lucas Riley. Our illustrator is Megan Goff, and our sound designers and engineers are Jake Cook and Jesse. Nice one, smaller. I'm Dina Temple Raston, and thanks for listening.

[Recorded Future Representative]

Support for this program comes from Recorded Future. In cybersecurity, the biggest risk isn't what can be seen, it's what gets missed. Recorded Future analyzes billions of signals to help organizations stay ahead of threats. Recorded Future Know what matters. Act first.

[Recorded Future News Announcer]

If you're looking for a daily guide to cybersecurity news and policy, sign up for the Cyber Daily from Recorded Future News. It serves up the day's most interesting and important cyber stories from our sister publication the Record, and then aggregates all of the big cyber stories you might have missed from news outlets around the world. Just go to the Record Media and click on Cyber Daily to get all you need to know about the world of cybersecurity right in your inbox.