By Dr. Randall Hanifen
“How will AI transform the fire-EMS service over the next five years?” Asking ChatGPT a question about its own technology provides some eye-opening answers.
Here’s a snapshot:
It should come as no surprise that AI will affect many aspects of our work, but how will funding and human interaction affect the pace of change? That’s what we’ll explore here.
| MORE: When the machine thinks for us
But first, AI would not be the people-pleaser it is if it did not suggest completing another task — in this case, developing a five-year AI transformation plan outlining the plan year, area of focus and key outcomes.
Here’s what it offered:
Let’s consider how this might play out in the areas of governance, dispatch and operations.
Building effective AI policy and governance
After sitting through an AI readiness course last year, I began to recognize the importance of AI governance, especially as it relates to government operations, specifically fire and EMS. While the typical user just adds prompts to the AI platform, many fire and EMS managers pay for more advanced solutions because they have access to more sensitive information, including medical data protected under federal privacy laws.
Because AI uses information fed into the system and is a people-pleasing tool known to contain inaccuracies, policies governing its use are important. Such policies ensure you are not violating federal law, betraying public trust, or committing a crime by sharing sensitive information with the platforms.
Building AI policies begins with understanding risk. The National Institute of Standards and Technology (NIST) has developed a framework — NIST AI Risk Management Framework (AI RMF) — to better manage AI-associated risk for individuals, organizations and society. Additionally, the U.S. Department of Homeland Security (DHS) has posted it principles for prohibitions for cases specific to DHS missions.
DHS established the following key principles for AI use:
- Lawful and mission-appropriate
- Mission-enhancing
- Safe, secure and responsible
- Trustworthy
- Human centered
DHS also outlines the following guidance and requirements for fulfilling its AI principles:
- Dynamic governance
- Human oversight
- Enterprise AI risk management
- Testing and evaluation
- Data management
- Responsible and authorized acquisition
- Workforce training
This list of guidance and requirements is likely one of the most critical components of AI policy.
While there is a race to enhance technology and be more business-oriented, fire and EMS leaders must fully evaluate the pros and cons of AI use. They must also be very specific about testing, the expected outcomes and when to “pull the plug.”
How AI is reshaping modern fire-EMS dispatch
Some of the AI programs under development or already developed will change dispatch as we know it today. I have seen technological developments in the past few years that allow dispatch centers to locate in areas completely remote from their service areas. This advancement could enable dispatch operations to be located overseas, as many businesses do by outsourcing call centers. It could also drive the consolidation of dispatch centers as the costs of operating these technologically advanced systems continue to rise.
With technology pinpointing caller location and text-to-voice dispatching, the need for local dispatchers lessens and the time needed to process calls decreases. Voice recognition and AI-based call triage will streamline and accelerate dispatch call-taking.
I believe predictive analytics is one of the most promising advancements for fire and EMS agencies. This technology supports pre-deployment resource allocation by analyzing multiple factors, including call history. While it may seem futuristic, these capabilities can be achieved through mathematical probability modeling or through commercial platforms such as FirstWatch, MARVLIS and LiveMUM.
One limitation, however, is that these systems are generally most reliable in large fire and EMS agencies; their accuracy and margin of error decline when the call volume used in the dataset is too small. Even so, these programs can help reposition units to cover high-demand areas and fill gaps left by units that are moved.
Emerging AI-driven tools for field operations
We are already seeing increasing interconnectedness among tools in the EMS field — for example, cardiac monitors that integrate directly with reporting software offer improved interpretation and provide expanded monitoring of multiple body systems. The next anticipated step is integrating AI into these devices to guide patient care.
With a patient’s history pulled from reporting systems and combined with body systems monitoring, it would not be a stretch to imagine AI-driven recommendations for EMTs and paramedics based on this data. However, this raises ethical concerns if the guidance provided is incorrect and providers rely on it in the field. For this reason, I believe full automation remains further from implementation.
Drones are increasingly being used to surveil incident scenes, providing thermal imaging and other critical data that enhance the incident commander’s ability to make informed decisions. An emerging capability in this area is the drone as first responder (DFR) model, in which drones deploy ahead of fire and EMS to provide critical, real-time information.
Companies such as Skydio, JOUAV and WISPR Systems are currently active in this space. In addition, the autonomous Drone Express platform from TELEGRID, capable of carrying up to 5 pounds, opens the possibility of remotely delivering an AED during a cardiac emergency.
Body-worn cameras, already common in police work, are likely to become standard for firefighters as well. Paired with breathing apparatus data, body systems monitoring and environmental sensors, these body-worn cameras could give incident commanders the ability to see inside a structure and assess real-time fire conditions. With these tools, smoke reading will no longer be the primary source of information from the command post.
Final thoughts
While technology has gradually advanced our profession over the past 25 years, the changes we can expect in the next five years are poised to outpace that progress by a factor of 10.
In our next installment, we will explore strategic planning, administration and training and safety.
| WATCH: Chad Crouse talks AI — We can’t ‘outsource all the thinking’
