Officer working a scenario in MILO VR and the computer-generated character he is interacting with. The AR option for MILO VR makes such scenarios more realistic.MILO
Teaching officers how to talk to people while wearing a uniform and a badge of authority has never been a quick process. It’s the kind of thing that comes with experience. Now, a new simulation feature from MILO Range may reduce the length of that communications learning curve.
MILO makes a wide variety of law enforcement training tools, including MILO VR, the company’s virtual reality law enforcement training solution and MILO Focus3D. MILO VR lets officers put on a headset and enter a computer-generated world that offers a variety of opportunities to interact with the types of people they will interact with on duty. The VR effect makes the scenario much more immersive for the trainee. And now a new MILO technology can make that experience more realistic.
At the 2024 International Association of Chiefs of Police Show, MILO showed a generative artificial intelligence feature for MILO VR that gives the computer-generated characters inside the simulator’s VR world some autonomy and a much more lifelike ability to speak with the officer. The instructor sets up the character, choosing from various different options in the system, then sets the character’s level of compliance and other attributes, including key phrases the character should say.
The AI responses of the characters in MILO VR and Focus3D are more natural than what a student experiences in a projection simulation or even in a VR scenario without the AI. “It’s a much more organic conversation,” says Lawrence Kehoe, a law enforcement subject matter expert at MILO.
Kehoe also points out a major advantage to the AI feature is that it helps the instructor focus more on the scenario’s training goals. “In a system without the AI, the instructor is worrying about how the character would answer the officer. With the AI, the instructor is paying more attention to the training and how the trainee is functioning within the scenario,” he says.
De-escalation training is an area where Kehoe also sees great value in the AI feature. “I think it will help demonstrate how well the trainee has learned how to de-escalate situations,” he says. Kehoe adds that the speed and variety of the character’s response will challenge the students more than instructor-generated responses.
Another training benefit of the AI is that each student’s experience in the scenario will be different based on the character’s response. “If you’re running five officers in a row through the scenario, the ones who have completed the scenario will not be able to tell the others what to expect. There’s absolutely no gaming of the system,” Kehoe says.
Kehoe believes the greatest benefit of adding AI to simulator training is the variety of responses the characters can make because it allows the instructor to see how the student responds in different situations. “It just opens up a lot more possibilities,” he says. “The best way to see how the trainee will react is to put them into a an environment where they have the freedom to make different decisions. The AI feature for MILO VR gives us a safe environment to do that.”
The AI feature for MILO VR and Focus3D is built from more than one large language library, say Sivan Segev, MILO’s head of research and development and chief developer of the software. However, it is not a typical AI tool because it has no access to the cloud or the Internet.
“For customer security reasons, we are not communicating with outside servers,” says Robert McCue, executive director of MILO. “It was designed as a closed loop.”
MILO’s generative AI is available now as an add-on feature in the MILO VR and Focus3D systems.
