I was sitting with two peers at the National Association of EMS Physicians conference this year when the conversation turned to a vendor working on a driverless ambulance.
At first, we laughed. Then we paused. Because we all realized something at the same time:
This isn’t science fiction anymore.
|WATCH NOW: FDNY’s future: AI, BWCs and pay parity with Commissioner Lillian Bonsignore
Artificial intelligence is already shaping dispatch systems, documentation, scheduling and clinical decision support. And like every new technology EMS has ever faced — from radios, to cardiac monitors, RSI, electronic PCRs, etc. — we have two choices:
We can understand it ... or be dragged along by it.
To lead our agencies responsibly, we need a simple way to think about AI maturity. One of the most useful frameworks is the five stages of AI development. It helps leaders understand what AI can do today, what’s coming next, and how to use it safely.
Let’s walk through it in EMS language.
Stage 1: Rule-based AI (protocol automation)
This is where most EMS systems already are.
Rule-based AI follows programmed logic. It doesn’t learn. It doesn’t reason. It just executes.
Examples in EMS include:
- CAD triage scripts
- Drug interaction alerts
- Protocol checklists
- Billing edits
- Phone triage decision trees
Think about your cardiac arrest checklist in CAD. It reminds crews of steps, but it doesn’t adapt.
- Strength: Rule-based AI is reliable and auditable.
- Weakness: It cannot adjust to real-world complexity.
This is the same kind of automation EMS adopted decades ago with protocol cards and dispatch algorithms.
Stage 2: Learning AI (predictive analytics)
This is where progressive EMS systems should be today.
Machine learning systems analyze large datasets to identify patterns.
Examples include:
- Predicting call volume by time and location
- Forecasting staffing needs
- Identifying high-risk frequent callers
- Detecting billing errors
- Predicting hospital offload delays
Research shows predictive deployment can improve response times by 5–15% when implemented properly.
- Strength: Learning AI can help you place units better, staff smarter and plan budgets with more accuracy.
- Weakness: It only works inside one task. It doesn’t “think.” It recognizes patterns.
Stage 3: Context-aware AI (assistive AI)
This is the stage we are entering now.
Generative and context-aware AI systems can process language, reports and complex information to assist humans.
Examples in EMS:
- Drafting PCR narratives from monitor data
- Summarizing QA reviews
- Creating training simulations
- Providing protocol explanations
- Supporting medical director case review
Imagine a new paramedic asking, “Why do we give nitro in pulmonary edema?” and an AI system pulling physiology, protocol guidance and teaching points instantly.
That’s assistive AI. It’s like having a knowledgeable preceptor: helpful, but still needs a senior medic watching.
- Strength: Context-aware AI improves education and workflow.
- Weakness: It needs human oversight.
Stage 4: Autonomous AI (operational decision systems)
This is where the driverless ambulance conversation starts to make sense.
Autonomous AI can plan and execute multi-step decisions with minimal supervision.
Potential EMS applications in the future include:
- Real-time system deployment adjustments
- Automatic scheduling and overtime management
- Supply chain optimization
- Automated interfacility transfer coordination
- Clinical triage routing
Imagine an AI system monitoring call demand, weather, staffing levels and hospital status ... and repositioning units automatically.
This could dramatically improve response reliability in rural systems.
But it raises serious questions:
- Who is accountable for decisions?
- How do we audit errors?
- What happens when the AI is wrong?
EMS has always operated under medical direction. AI must operate under governance.
- Strength: Autonomous AI could improve system efficiency, deployment and resource management at a scale humans cannot match in real time.
- Weakness: It could create accountability, safety and oversight challenges when systems make operational decisions independently.
Stage 5: Artificial general intelligence (AGI)
This stage has not been reached.
AGI would be able to reason across domains like a human clinician or executive.
It could theoretically:
- Diagnose complex cases
- Design EMS systems
- Write protocols
- Manage agencies
- Teach leadership
But we are not there today. And responsible leaders should separate hype from reality.
- Strength: AGI could eventually combine clinical reasoning, operational planning and leadership support into a single system.
- Weakness: No one fully understands the risks, governance challenges or unintended consequences of systems operating at that level of intelligence.
Why AI innovation matters for EMS leaders
Every major EMS innovation faced resistance:
- Cardiac monitors
- Electronic PCRs
- RSI
- Telemedicine
Yet the systems that adapted early improved patient care and workforce stability. AI is the same. The question isn’t whether AI will enter EMS. It already has. The real question is whether EMS leaders will guide it responsibly.
Practical steps for adopting AI in EMS today
1. Start with low-risk applications
Use AI where mistakes are easily caught:
- Documentation support
- QA summaries
- Training simulations
- Scheduling optimization
These save time without risking patient care.
2. Build governance before deployment
AI should operate under:
- Medical director oversight
- QA audit trails
- Clear accountability policies
- Data privacy protections
If you wouldn’t trust a new medic without supervision, don’t trust AI without oversight.
3. Train your workforce
Fear comes from misunderstanding.
Teach crews:
- What AI can do
- What it cannot do
- How to verify its outputs
This builds trust and prevents misuse.
4. Protect clinical judgment
AI should support medics, not replace them.
EMS is still about:
- Scene awareness
- Patient empathy
- Ethical decisions
- Clinical reasoning
No algorithm can replace the human moment of patient care.
5. Think long-term
Driverless ambulances may or may not become practical. But AI-assisted EMS systems absolutely will. Leaders who prepare now will protect their patients, their crews, and their agencies.
A final thought
That conversation at NAEMSP stayed with me, not because driverless ambulances are around the corner, but because it reminded me of something important.
Technology doesn’t define EMS. Leadership does.
If we understand AI, we can shape it to improve response times, reduce burnout, strengthen education and deliver better patient care.
If we ignore it, someone else will shape it for us.
And EMS has never done well letting outsiders decide how we practice.
The future is coming.
Let’s meet it with knowledge, ethics, and the same professionalism that has carried this field forward for 50 years.
