> ## Content Index > Fetch the complete content index at: https://www.holdtraffic.com/llms.txt > Use this file to discover other available public pages before exploring further. # AI Has Entered the Chat. - URL: https://www.holdtraffic.com/ai-has-entered-the-chat/ - Published: 2026-09-13T14:30:32.000Z - Updated: 2026-09-13T14:30:40.000Z - Description: AI isn't coming to the 911 center. It's already here. The question isn't whether we use it—it's where we use it without losing the human judgment that makes emergency communications work. - Author: Justin Storch - Tags: 911dispatch, tech, News, public safety communications Artificial intelligence has officially entered the 911 center. Depending on who you ask, that statement is either really exciting, downright terrifying, or somewhere in the middle. I recently wrote a short piece about Jonesboro 911 in Arkansas, testing an AI system to answer non-emergency calls. That got me to looking at the bigger conversation around AI in public safety communications. Let's be honest, answering the non-emergency line is just the tip of the iceberg. AI is already being utilized for call triage, transcription, translation, quality assurance, training, data analysis and more tasks than you can shake a stick at. As Next Gen 911 continues to pull more information into the center, the amount of data that we are asking telecommunicators to process is only going to increase. This is where I think we need to change the conversation a little. We keep asking: **Will AI replace dispatchers?** My personal opinion: No. But I'm not convinced that is even the question we should be asking. Maybe the better question is: **What actually needs a dispatcher?** Now hold on before you start throwing the proverbial rotten produce at me, hear me out. That sounds harsher than I intend it. A telecommunicators value isn't in their ability to look up a phone number, answer what time the animal shelter opens, or answer the same administrative question for the twentieth time that shift. The value is in their judgement. It's hearing something in a caller's voice that doesn't sound just quite right. It's the ability to recognize that the "civil matter" being described may actually be domestic violence. It's knowing when an officer's radio traffic suddenly sounds different. It's calming the parent whose child isn't breathing while getting responders headed their way. It's understanding context, emotion, danger and uncertainty and then making decisions when the information coming through the headset doesn't fit neatly into a box. That's the part of 911 we should be protecting. So instead of asking ***whether*** AI belongs maybe we should be asking ***where*** it belongs. ### **AI Is Already Here** This isn't a conversation about some distant future. ![](https://static.klipy.com/ii/c3a19a0b747a76e98651f2b9a3cca5ff/52/12/NfS3AJwk.gif) Documentation exists on uses of AI in emergency communications, including call triage, transcription and translation. NTIA (The National Telecommunications and Information Administration) reported that one AI call triage implementation processed 9,635 calls during April of 2024\. Of those 2,920 (about 30%) were identified as non-emergency or general information calls and resolved without requiring a call taker. Think about that from the console.... That's not 2,920 calls where a computer "replaced a dispatcher". That's potentially 2,920 interruptions a dispatcher didn't have to handle. And anyone who has spent a busy day on a console that just won't quit understands that difference. The problem in communications centers isn't always that we don't have enough seconds in a day. Sometimes it's that we are trying to give those same seconds to five different things. This may be where AI becomes valuable. Not by replacing the person wearing the headset, but by taking some of the noise out of it. ### Where Could AI Actually Help? **Non-Emergency Calls** We've already talked about this one. Routine questions, administrative calls and request for basic information consume attention. If an AI system can reliably handle those calls while immediately recognizing when something needs to be transferred to a telecommunicator, there is an obvious potential benefit. But that last part is critical The system has to recognize when: *"I just need an officer to talk to my husband..."* isn't actually the routine call if first appeared to be. **Transcription** This may be one of the least controversial applications. Imagine having a searchable, accurate transcript available when you're handling a call. The caller rattles off a license plate, a complicated street name, subject description, or a phone number. Instead of having to ask the caller to repeat or relying entirely on memory, the information is on the screen. That's not replacing judgment. It's supporting it. **Translation** Language barriers are another area where AI could and is making an enormous difference. Traditional language line services work, but they introduce another person and another step into a conversation where seconds matter. Real time AI translation could potentially allow telecommunicators to communicate directly. Again, AI is not making the decisions. It's removing the barrier between the person making the decision and the person who needs help. **Quality Assurance** This one is changing how we approach QA. Most centers can only review a fraction of their calls. AI potentially makes it possible to analyze far more of them. Was the address verified? Were required questions asked? Are there patterns developing across certain call types? NENA has previously discussed AI's potential to review calls and identify trends that could improve training. And unless you have been living under a rock, you have probably seen the ever-growing list of AI assisted QA programs. But this is also an area where leadership matters enormously. There is a huge difference between using AI to identify opportunities for improvement and creating an automated disciplinary machine watching every word a dispatcher says. ![](https://static.klipy.com/ii/d6b0ce929193df3c242ac34b5654d2ce/a9/c1/f9c4AsCy.gif) If employees believe AI exists primarily to catch them doing something wrong...you've probably already lost the room. **Information Overload** This may eventually become the biggest argument for AI in 911\. Next Gen 911 isn't just bringing us better telephone calls. It's bringing data. Text, images, video, telematics, sensor information, automatic crash notification, and whatever comes next. Guess what.... someone or something has to sort through all of that. The dispatcher cannot reasonably stare at six additional screens while listening to two radios, talking to a caller and entering information into CAD. At some point we have someone has to decide what information actually deserves the telecommunicator's attention. And maybe that's the role AI was built for. Not making decisions. **Helping the human see what they need in order to make the decisions.** ### **But There Are Some Big Questions** This is where my enthusiasm starts to get a little more cautious. Because 911 isn't a customer service center. If an AI chatbot for a retailer misunderstands someone, maybe the customer gets frustrated. If an AI system in a 911 center misunderstands someone, the consequences can be considerably higher. What happens when AI is confidently wrong? What happens when it misunderstands an accent? Or translates something wrong because it's a specific dialect? What happens when the network goes down? What information is being stored? Who has access to it? How is CJIS protected information handled? Who is liable when it fails? And most importantly.... Does the telecommunicator still feel empowered to tell the computer it is wrong? That question matters more than it may seem. Technology has a way of becoming authoritative simply because it's on the screen. If the computer says something with enough confidence, eventually people stop questioning it. That cannot be allowed to happen in 911\. AI should inform human judgment. It should never replace the expectation that the person at the console continues to think. ### Don't Automate the Humanity Out of 911 There is another danger that doesn't involve cybersecurity, liability or bad algorithms. We could simply become too efficient. Emergency communications is one of the rare places where government meets people during some of the worst moments of their lives. The person answering the phone isn't merely gathering data. Sometimes they are reassuring someone. Sometimes they are listening. Sometimes they're creating order out of absolute chaos. Sometimes the most important thing happening on that call isn't something that belongs in a CAD field. We shouldn't automate that away simply because we can. There will be places where AI makes tremendous sense. There will be places where it doesn't belong at all. Figuring out the difference may very well be one of the most important technology conversations our industry has over the next few years. ### The Dispatcher Isn't the Problem We're Trying to Solve I don't think the goal of AI in 911 should ever be figuring out how to operate a communications center without telecommunicators. The goal should be figuring out how to build a communications center where telecommunicators can spend more of their attention doing things only they can do. Let the computer transcribe. Let it organize information. Let it identify patterns. Let it answer routine questions. Let it manage the flood of data headed for the modern ECC. But when judgment matters... When empathy matters... When something doesn't sound right... When someone's worst day is happening on the other end of the phone... **Keep the human in the chair.** ### Sources and Further Reading Want to dig a little deeper into AI and the future of emergency communications? Here are a few of the resources used in putting this article together: **National Telecommunications and Information Administration (NTIA)** [*AI-Driven Transformation in 9-1-1 Operations*](https://www.ntia.gov/other-publication/2025/ai-driven-transformation-9-1-1-operations?utm%5Fsource=chatgpt.com) A look at how AI is being used and explored in emergency communications, including call triage, transcription, translation, and other applications. This is also the source for the call-triage example cited above, where 9,635 calls were processed and 2,920 were resolved without requiring a call taker. **APCO International** [*Artificial Intelligence in Emergency Communications*](https://www.apcointl.org/technology/ai-in-emergency-response/?utm%5Fsource=chatgpt.com) APCO's resources on the use of AI in ECCs, including operational considerations, cybersecurity, ethics, training, and the development of industry best practices. **NENA: The 9-1-1 Association** [*NG9-1-1 Operational Impacts on the PSAP*](https://www.nena.org/resource/resmgr/standards/nena-inf-041.1-2021%5Fng911%5Fop.pdf?utm%5Fsource=chatgpt.com) NENA's discussion of the operational impact of Next Generation 911, including the increasing amount of data coming into emergency communications centers and the role technology and AI may play in helping telecommunicators manage it. **KAIT8 — Jonesboro, Arkansas** [*Jonesboro Dispatch Center Is Testing an AI Call Taker*](https://www.kait8.com/2026/09/02/jonesboro-dispatch-center-is-testing-an-ai-call-taker/?utm%5Fsource=chatgpt.com) A look at Jonesboro's implementation of "Lezle," an AI system being tested to handle non-emergency calls and reduce the workload placed on human telecommunicators.