More Than Just a Body Camera: AI Integration Adds Enhanced Features for Law Enforcement

Axon Assistant

During Axon Week 2025 in Phoenix, Arizona, in April, Rick Smith, CEO of Axon, introduced the company’s new voice-enabled AI companion, Axon Assistant, which is built directly into the Axon Body 4 body worn camera. “Axon body cameras was the moment that transitioned us from a weapon company to a much broader ecosystem. Like our software platforms, body cameras have become more critical and more valuable with AI,” he told the audience. 

“Two years ago, we launched our latest body worn camera, the Axon Body 4, which for the first time began to allow two-way communication. Now we've added AI to the mix. What was once just a recording device is now a completely connected communication support tool. It's sitting at the core of the officer and able to do so much more than we ever imagined when we created it.”

Upon launch, Axon Assistant includes three key skills: Real-Time Translation, which supports over 50 languages; Voice-Enabled Policy Chat, which allows officers to ask questions and get clear, immediate answers to agency policy questions; and General Q&A, which gives officers the ability to ask Axon Assistant everyday questions on facts that could be useful for them in the field. One of the main goals of Axon Assistant is to help the officer keep their eyes on a scene without having to turn away for information. More skills, including license plate lookup, weather and traffic information, are expected to launch later in 2025.

Following the announcement, Jeff Kunins, Chief Product Officer and Chief Technology Officer at Axon, spoke about his excitement for Axon Assistant, having spent much of his career focusing on AI voice assistants. Before joining Axon in 2019, he was Vice President of Amazon’s Alexa Entertainment and before that was VP of Product at Tellme Networks through their acquisition by Microsoft in 2007. He says that back then, the timing and technology wasn’t right in a way that would be helpful to officers. “I actually resisted it back then. But now is the time.”

Currently, the Axon Assistant feature is activated by push-to-talk. “We really like the push to talk in the context of the body camera because you're so often in loud, noisy environments. We want to make sure the officer feels in control,” he says. “It also has familiarity relative to the radio. Another thing to think about is that we want to normalize for officers the habit of using the body camera for communications in general. Talking with the assistant is a great way to get started.” 

With two-way audio for live-streaming already included in the Body 4 software, he says that introducing officers to skills that harness the power of AI will come with time. “As we add more capability into the camera, we want them to feel comfortable on an all day, everyday basis, just communicating with it.”

Motorola Solutions Assist

While Motorola Solutions announced its SVX Video Remote Speaker Microphone and its own AI companion “Assist” in April, it spotlighted the new technology and rolled out the Assist Chat feature during the Motorola Solutions Summit 2025 in Grapevine, Texas, in May.

The SVX video remote speaker microphone combines video, voice and AI in one device and serves to augment the company’s APX NEXT radio. Todd Piett, Senior Vice President of Command Center at Motorola Solutions, spoke about the SVX, Assist and how the company plans to make an officer’s shift easier through technology. "I'm not a hardware guy, so when I first heard how we were building a new integrated microphone and video device, I said, ‘OK, that's kind of cool.’ ” Then he saw an initial demo of the device in action. “It really is amazing. It combines two devices and AI. This device allows you, without compromise, to have the same voice quality all the Motorola radio microphones have.”

With the addition of high-quality video, as well as AI and processing at the edge, the device comes with a variety of useful built-in capabilities. One thing featured in the SVX is the AI-assisted voice quality. “The thing that stood out most, and I just didn't really appreciate it until I saw these demos, was the voice quality. When this thing is activated, you can whisper hold it down low. It’s drowning out all of the ambient noise. It sounds clear as day over the radio, and it’s still recording all the ambient noise for evidentiary workflow.”

Advances in both voice transcription and translation made Assist possible, according to Piett. “Previously, best in class was maybe 70% to 80% transcription accuracy off of this and we're getting well over 90%,” he says. “That’s the difference between making an error and being able to automate something. So, we're super, super excited.”

The AI in Assist doesn’t only function in the SVX, but carries through Motorola Solutions’ ecosystem of products. “We’re capturing audio for both the radio, for the radio traffic. We're capturing video and audio traffic from this. We have all of the CAD and incident data from dispatches. We have the 911 call audio. We’re carrying that all the way through the workflow into the narrative creation for the officers,” he says. “Each of those data sets can be used as a proof point to verify that what I'm storing is a record and when I'm creating evidence is actually what the officer saw.”

Chris Bennett, Director of AI Transparency and Education at Motorola Solutions, coverage the implications of AI extensively at the annual conference, stressing that even if the introduction of AI only saves officers a matter of seconds in any given incident, it will help save lives.

“Nobody wants AI to sprinkle on everything because it's newer, because it's popular. It can become a distraction. You don't want to be bombarded with too many different options. What we want is for the existing applications that you use just become more intuitive, more friendly, more helpful to you,” he says. “You don't need training to know what button to hit on the side of the radio. You shouldn't need training to know how to leverage AI in your application. You see headlines about how AI saved X number of hours in this situation or weeks or months. That's not our overall goal. Of course, that's the end goal, but that goal is achieved by saving seconds or minutes in each different step of the process.”

Bennett says that in any given emergency call, there may be dozens of people involved. “They have buttons to click; they have things to send to each other. They have things to listen to. If you could just shave seconds off of each one of those parts of the process without requiring them to go through any training or learning new technology, but you just use AI to make all of that faster.” 

Bennett says that Motorola has been involved in artificial intelligence, pointing out that in some of the early object recognition in its video or cybersecurity functions for threat analysis and analytics. Even the 911 call transcription used AI. “With the SVX, the biggest differentiator there is just how good the audio quality is. You can have lights and sirens and all kinds of stuff going on. Say that I'm speaking with a lieutenant here and all this other stuff is going on, but I’ve got this crystal-clear audio. Not only do I have that crystal-clear audio that I can then channel in and do an AI analysis on the transcription, but then you also have the evidentiary audio that can be used in court.”