How AI Is Changing 911 Centers: From Non-Emergency Overload to Smarter Emergency Response

Every day, over 240 million calls are placed to 911 centers across the United States each year.

But here’s a number that might surprise you: a significant portion of those calls are not true

emergencies. Storm damage reports, graffiti complaints, towing inquiries, noise complaints, and

general information requests flood Public Safety Answering Points (PSAPs), pulling trained

telecommunicators away from the calls that truly require their expertise and urgency.

At the same time, America's 911 workforce is in crisis. According to a national survey by the

International Academies of Emergency Dispatch (IAED) and the National Association of State 911 Administrators (NASNA), the average vacancy rate in 911 centers across America was

approximately 25% between 2019 and 2022. That means one in every four positions needed to

be filled. In some centers, the vacancy rate exceeded 70%.

This combination of rising call volumes and shrinking workforces has created an urgent need for

new solutions. And in 2025 and 2026, those solutions are arriving in the form of artificial

Intelligence (AI).

AI is not replacing 911 dispatchers. And in this blog from NGA, we’ll discuss how artificial intelligence is instead giving them something they desperately need: relief from the routine, so they can focus on the life-threatening calls that demand human judgment, empathy, and split-second decision-making. 

The Non-Emergency Overload Problem

To understand why AI matters for 911, you first need to understand the scale of the non-

emergency problem.

Studies from PSAPs across the country show that a large percentage of incoming calls are not

actual emergencies. Arlington County, Virginia, for example, found that a substantial portion of

their call volume consisted of non-emergency requests: towing inquiries, storm damage reports,

abandoned vehicle complaints, and general information calls. These are important community

needs, but they don’t require a trained emergency dispatcher to handle.

Every minute a telecommunicator spends on a routine call is a minute they’re unavailable for a

cardiac arrest, a structure fire, or an active threat situation. When staffing is already stretched

thin, these non-emergency calls create dangerous bottlenecks.

The National Telecommunicator Institute (NTIA) published a white paper in November 2025

titled "AI-Driven Transformation in 9-1-1 Operations" that outlined the scope of this challenge

and the AI solutions already being deployed to address it.

What AI Is Actually Doing Inside 911 Centers Today

AI in 911 isn’t a theoretical concept or a future promise. It’s already deployed and delivering

measurable results at PSAPs across the United States. 

Here are some of the real-world applications:

Non-Emergency Call Routing and Triage

AI systems can identify non-emergency calls and automatically route them to appropriate

resources. This includes directing callers to online portals, self-service tools, or non-emergency

hotlines without ever requiring a human dispatcher to intervene.

• AI reduced non-emergency call volume by up to 40%, freeing telecommunicators to focus on genuine emergencies. Jeffcom 911 (Jefferson County, Colorado)
• AI reduced non-emergency calls by approximately 36% on average, with a peak near 39% in 2024. (Monterey County ECD California)
• AI triage averages more than 40 events per day and has enabled up to 40 seconds faster answer times for other callers on average. The center has achieved near-zero downtime, with only 2.5 hours of downtime since June 2022. Orleans Parish Communications District (New Orleans)

Real-Time Transcription and Call Summarization

Natural Language Processing (NLP) and Large Language Models (LLMs) are being used to

transcribe 911 calls in real time, detect critical keywords, analyze caller sentiment, and generate

incident summaries. This means dispatchers can review incoming information faster and pass

accurate details to first responders without manually typing notes during a high-stress call.

Language Translation

AI-powered translation tools now support real-time interpretation in hundreds of languages. For

PSAPs serving diverse communities, this eliminates the delay of connecting to a human

interpreter. AI-driven translation tools already implemented into state-of-the-art emergency systems, like NG911 (Next Generation 911), connect non-English-speaking callers to emergency services in seconds rather than minutes.

Predictive Staffing and Resource Allocation

AI can analyze historical call data, weather patterns, local events, and other variables to forecast call volume surges before they happen. This allows PSAP managers to adjust staffing proactively rather than reactively scrambling when call queues spike. 

Real-World Results: Arlington County Sets a National Precedent

One of the most widely recognized AI deployments in 911 is at the Arlington County, Virginia,

Emergency Communications Center (ECC). In 2025, Arlington County was honored with an AWS Champions Award for its implementation of Amazon Connect to modernize 911 call handling.

Arlington County's cloud-based system uses AI and automated voice response to manage non-

emergency calls, including storm damage reports, graffiti complaints, and towing inquiries. The

AI intelligently routes calls, responds to common inquiries, and redirects callers to appropriate

digital self-service resources.

Critically, the system preserves human call takers for high-priority emergency situations.

Arlington County's approach demonstrates what responsible AI adoption looks like in public

safety: automation handles the routine, and skilled human beings handle the critical.

The results speak for themselves: reduced hold times, faster service for non-emergency callers,

improved operational efficiency, and a better experience for callers who reach a live dispatcher

when they truly need one.

The Human-in-the-Loop Principle: Why AI Will Not Replace 911 Dispatchers

One of the most common questions in the public safety community right now is: "Will AI replace

911 dispatchers?" The answer is no, and here is why.

A 911 telecommunicator does far more than answer a phone. They assess chaotic, high-emotion situations. They make split-second triage decisions that determine whether police, fire, or EMS responds. They provide pre-arrival instructions that can save a life, such as guiding a bystander through CPR. They calm panicked callers who cannot clearly communicate what is happening.

These are fundamentally human skills that AI cannot replicate. What AI can do is remove the

non-emergency burden, handle the repetitive administrative work, and provide real-time data

support so that dispatchers can perform their critical role more effectively. The NTIA's 2025 white paper summarized it well: "AI fundamentally supports telecommunicators by reducing repetitive tasks and enhancing real-time information access, with the potential to alleviate shortages and burnout."

This is the human-in-the-loop model. AI handles the noise. Humans handle the emergencies.

Together, they create a system that is faster, more efficient, and more sustainable for the

people who do one of the hardest jobs in public safety.

Why Cloud-Native NG911 Infrastructure Makes AI Possible

There is an important technical foundation behind all of this: AI in 911 requires modern, cloud-

based NG911 infrastructure to function properly. 

Legacy 911 systems, which rely on analog, circuit-switched networks, were not designed to

support AI workloads. They cannot process the data volumes, run the machine learning models,

or integrate with the third-party AI tools that PSAPs are now adopting.

Next Generation 911 systems built on IP-based Emergency Services IP Networks (ESInets) provide the digital foundation that AI requires. Cloud-native platforms enable scalable call handling, rapid deployment of new tools, and disaster resilience. They support text, video, photos, and data alongside traditional voice calls, giving AI models richer input to work with.

As the NTIA white paper noted, "Transition to cloud-native platforms is fundamental for AI,

enabling scalable call handling, rapid tool deployment, and disaster resilience." Orleans Parish

Communications District, one of the leading AI adopters, has demonstrated exactly this on their

cloud-based NG911 system.

For PSAPs considering AI adoption, the first step is ensuring that their underlying NG911

infrastructure can support it. The platform matters as much as the AI tools themselves.

What PSAPs Should Consider Before Adopting AI

AI is not a plug-and-play solution. 

PSAPs that are evaluating AI tools should consider the following:

• Start with the biggest pain point. If non-emergency call volume is overwhelming your team, AI call triage and routing should be the first deployment. If language barriers are a challenge, AI translation tools offer immediate impact. 
• Ensure your NG911 platform supports AI integration.Cloud-native, IP-based infrastructure is a prerequisite. Legacy systems will need to be modernized first. 
• Demand transparency from AI vendors. Ask how their models are trained, what data they use, how they handle bias, and what safeguards are in place for misclassification. A non-emergency call misrouted as "low priority" when it is actually a medical emergency is unacceptable. 
• Prioritize cybersecurity. AI tools that integrate with your NG911 system must meet the same security standards as the system itself. Ensure vendors provide end-to-end encryption, access controls, and compliance with NENA standards. 
• Keep humans in the loop. Every AI deployment should preserve a clear path for a human dispatcher to take over at any point. Emergency calls must always reach a trained telecommunicator.

The Future of 911 Is Human + A

The 911 staffing crisis is not going away. Call volumes continue to rise. Community expectations

continue to grow. And the telecommunicators who dedicate their careers to public safety

deserve better tools to do their jobs.

AI is not the answer to every problem in public safety. But when deployed responsibly, on

modern NG911 infrastructure, with a clear commitment to keeping humans in the loop, it is one

of the most promising developments the industry has seen in decades.

The PSAPs that are leading the way, from Arlington County to Jefferson County to New Orleans, are proving that AI and human expertise are not competing forces. They are complementary ones. And for the communities they serve, that combination is making emergency response faster, smarter, and more reliable.

Conclusion 

Is your PSAP ready for AI? It starts with the infrastructure. 

Contact NGA to learn more about their NG911 platform and how it provides the foundation for the next generation of emergency communications technology.