Hyper-turbulent times: EMS economics and AI guardrails with Matt Zavadsky and Dr. Shannon Gollnick
Recorded on the floor of the EMS|MC EMSpire Conference in Charleston, South Carolina, this episode of EMS One-Stop finds host Rob Lawrence in conversation with long-time collaborator and EMS advocate Matt Zavadsky.
Fresh off the longest federal government shutdown in history, Rob and Matt unpack what the hyper-turbulence in Washington really means for EMS: suspended Medicare extenders, disrupted grant programs, agencies taking out loans just to meet payroll and training programs put on hold.
They break down NAEMT’s flash poll on the shutdown’s impact, the promise of the Treatment in Place (TIP) legislation, and why associations “hunting as a pack” on Capitol Hill matters more than ever.
Along the way, they spotlight EMSIntel.org as a national barometer of EMS funding, staffing and response time crises, and issue a clear call to action for providers, billers and leaders to use association tools to contact their members of Congress.
| MORE: Government reopens: What EMS providers need to know right now
In the second half, Rob is joined by Dr. Shannon Gollnick, paramedic, EMS leader and organizational psychologist, to explore how artificial intelligence is reshaping EMS — right now.
Shannon makes the case that AI is “not the future; it is the present,” and that agency leaders must urgently build literacy, policies and guardrails around its use. They dig into the difference between HIPAA-compliant, embedded AI in ePCR systems, and risky open tools like ChatGPT, touching on hallucinations, embedded code and emerging Medicare fraud-detection programs.
| MORE: Artificial to augmented intelligence. How Dr. Shannon Gollnick wants EMS to work smarter, not harder
Rob and Shannon talk about AI as a powerful but potentially dangerous tool — “like having a tiger” — and outline practical steps for chiefs:
Ask: “Do we have an AI policy?”
Define what AI can and cannot be used for
Insist that every AI-generated work product is double-checked by a human before it hits the record
Memorable quotes
“We weren't here to actually scare you off it. We're here to let you know that it's here, but it's like having a tiger, right? We all love to have a tiger, but it has to be contained in some sort of guard, otherwise it's going to run rife and cause havoc, and we don't want that.” — Rob Lawrence
“This is part of the hyper-turbulence that's occurring in EMS right now.” — Matt Zavadsky
“So I think the message for the profession right now is, now is not the time to put your foot on the brake. It's time to put your foot on the gas.” — Matt Zavadsky
“We put the fun into function.” — Dr. Shannon Gollnick
“I think it's important to understand that AI is not the future. It is the present. We are currently here right now. And it's nothing to be afraid of.” — Dr. Shannon Gollnick
“If you're not doing it, I promise you that your staff is doing it and they're playing around with AI.” — Dr. Shannon Gollnick
“Guardrails don't exist from a congressional standpoint. They don't exist from a regulatory standpoint. The technology is moving far too fast. So we as agency leaders have to take the lead in putting up some of those guardrails.” — Dr. Shannon Gollnick
“There are ePCR software out there that are using proprietary AI that will use AI-generated narratives. And that absolutely is 100% good to go. What we don't want to see is our crews putting in their ChatGPT to have ChatGPT write their narrative.” — Dr. Shannon Gollnick
“ChatGPT has embedded code inside of it that you can't see, but that code is there ... so what we're kind of afraid to do is to say, hey, what happens 6 months from now, 8 months from now when Medicare does an audit, they run your ePCRs and find all of this embedded code from ChatGPT ... you open yourself up for a lot of compliance issues.” — Dr. Shannon Gollnick
Additional resources:
EMS Intel EMS News Tracker
American Ambulance Association Advocacy
NAEMT Advocacy
EMS shutdown survival: What leaders need to know now
Charting the future: How AI is rewriting the EMS narrative
Episode timeline:
00:21 – Rob introduces guest Matt Zavadsky
02:02 – Rob recaps the 40-plus-day federal government shutdown, questions about reopening, and his upcoming return to Capitol Hill for renewed advocacy
02:02 – Matt frames the shutdown as part of the “hyper turbulence” in EMS; explains the regulatory suspensions, pauses in Medicare extenders and grants, and how cash-flow uncertainty forced some agencies to take out loans just to make payroll
03:04 – Matt details NAEMT’s flash poll (408 agency responses) showing suspended training and grant-funded programs, and warns of a possible repeat shutdown around January 30
03:54 – Rob and Matt discuss the reopening of government, ongoing bipartisan work, and the risk that everything “comes to a grinding halt” again if Congress can’t agree
04:51 – Matt explains why NAEMT released the shutdown-impact poll even as government reopened and stresses the need to keep pushing for permanent relief from Medicare extenders and advancement of key bills like Treatment in Place (TIP)
06:03 – Matt outlines the House and Senate TIP companion bills and why Medicare paying for treatment in place is better for patients, EMS, the health system and the Medicare trust fund
06:54 – Rob notes broad association/provider support and professional lobbyists on the Hill; Matt stresses that field providers, administrators and billers must still use association legislative portals to send letters to Congress
08:08 – Matt describes a surge in communities reevaluating their EMS delivery models because of staffing, finance and subsidy challenges — “a great time to be an EMS consultant”
09:09 – Rob introduces EMSIntel.org as a curated clearinghouse of EMS news, used to show communities they aren’t alone; describes failed tax measures and funding referenda
10:15 – Matt cites EMS Intel data: ~85% of stories each month involve funding, staffing or response times; Rob and Matt stress the ubiquity of these themes from big cities to small towns
11:09 – Rob highlights mutual aid tensions and taxpayers questioning why they “pay to send our resources somewhere else;” both emphasize that hyper-turbulence and funding gaps are national issues
13:23 – Rob resets the scene from the EMSpire conference and recaps Matt’s Hill update before introducing Dr. Shannon Gollnick
14:41 – Shannon gives his backstory: in EMS since 1996, paramedic since 2002, progression into EMS leadership, doctorate in organizational psychology and focus on how organizations function
15:14 – “We put the fun into function.”
15:24 – Rob invites Shannon to talk AI, calling it “the specter we are embracing everywhere,” and references HIPAA concerns; Shannon opens with the core message: AI is not the future, it’s the present, and nothing to be afraid of
16:03 – Shannon urges leaders to build AI literacy, noting that if agencies aren’t using it, their staff and the younger generation already are
16:28 – Shannon emphasizes policy and procedure: AI guardrails aren’t coming from Congress or regulators, so agency leaders must define how AI will be used and where its limits are
16:55 – Rob reminds listeners that AI in EMS isn’t new, citing early monitor rhythm interpretation in the UK; Shannon underscores that crews already use AI tools and that unmanaged cut-and-paste practices can create billing and compliance risks
17:24 – Shannon explains the dangers of using open tools like ChatGPT for ePCR narratives: potential PHI exposure in a “black box” system and AI hallucinations generating plausible but false patient information
18:21 – Shannon describes how AI “wants to answer your question and make you happy,” leading to made-up details, and shares examples from testing minimal-input scenarios that returned overly detailed, inaccurate narratives.
19:03 – Shannon calls ChatGPT “kind of a snitch,” explaining embedded code markers that fraud detection tools — and increasingly Medicare’s AI-based “Wiser” program — can use to identify AI-written content in documentation
19:59 – Shannon warns about retrospective audits and compliance exposure if ChatGPT-coded narratives are found in ePCRs, noting that AI rules are still emerging and tech is outrunning regulation
20:51 – Rob summarizes the mixed message: AI is here and being built into devices and software, but there are real dangers. They discuss data going “to the cloud” — which Shannon defines as “somebody else’s computer.”
21:24 – Shannon frames AI as a powerful tool that can “put a lot of holes in the wall” if misused; he references fraudulent AI uses and deepfakes as emerging issues
22:05 – Shannon compares AI’s impact to the internet’s paradigm shift; Rob gives a “spoiler alert” about his own workflow using transcripts and ChatGPT agents, and notes the importance of reading and checking any AI-generated output
22:45 – Shannon reinforces that AI makes mistakes and cannot understand human context; he uses his “How you doing?” Joey Tribbiani vs. Tony Soprano example to illustrate contextual nuance
23:06 – Rob expands the context point with the “Friends”/“Sopranos” slide and reminds listeners that once AI-written words are published, “you said it.” Shannon highlights the WebMD effect and AI-driven self-diagnosis risks.
24:02 – They note that ChatGPT can generate long, complex diagnoses without sufficient patient context, leading to errant or misleading outcomes if misused clinically
25:00 – Rob summarizes: AI is here and, used correctly, is a good thing; advises chiefs to ask their teams, “Do we have an AI policy?”
25:27 – Shannon outlines what an AI policy should contain: acknowledgment that AI is here; clear, non-fearful framing; specificity on what decisions AI can support; and clarity on which tools (e.g., embedded EPCR AI) are allowed versus prohibited uses of ChatGPT
26:17 – Shannon stresses AI should not be used for clinical decision-making or clinical narrative writing; its role should be administrative only, and all outputs must be double-checked
Enjoying the show? Email
[email protected] to share feedback or suggest a guest for a future episode.
USA