I think that the most powerful use of AI ems is actually a passive listening feature that makes suggestions. Not necessarily tablet based or a PCR, but more of an Alexa that listens in and make suggestions as appropriate. Eventually, something like Star Trek holographic Dr. would be absolutely incredible from a patient care perspective- but that would require AI to be wise enough to build a story from extremely erratic and inconsistent elements.
While RocketMedic is currently taking a break from this forum...
I don't see AI being "passive" at all if it is listening and
then offering suggestions. To do so requires it to be an
active participant. With the level of AI capability right now, having it be something like the Star Trek EMH would take an astoundingly high level of programming and computing power to do. I can see AI having a "scribe" role in that it very well could take notes as an event progresses. You simply do as the practical exams require to you do: vocalize your actions as you do it. The AI can then present the transcription of events to you and you then choose what points to include in your charting. This would also allow the AI to be able to backfill part of your charts as needed. The ultimate review for this should be on you however. Because you are the one that are doing the charting, you are responsible for the ultimate completion of the chart.
I use voice dictation software at different times, as a case in point, I'm actually using the built in voice dictation software from Windows 11. When I am using my iPad, I use that built in dictation software to enter my narrative text into the EMR program that I am required to use. For the most part I actually generally use a preset template for my narratives because then I only have to enter the specific elements that are unique to that particular patient. When I do dictate an entire narrative, it is often because I am either bored or the particular patient is so generic that it would be easier for me to simply dictate the narrative rather than modify an existing template.
One of the instances in which I would use voice dictation is when I have to enter a lot of text very quickly. I can dictate my narrative at a rate of about 150 words a minute rather than typing at a rate of about 60 words a minute. I used to type slower than that but now I am much faster because, unfortunately, it's a side effect of having to type a lot of papers over the past few years. I first learned to type when I was much younger than I am now... if I put my fingers on the wrong keys when I start typing, the incomprehensibility is astounding.
As I stated earlier about the usage of AI in a quality check type situation, the AI could be used to verify or suggest various protocol options that might have been chosen, or possibly present to the medic a number of differential diagnoses that could be done in the field. That would be more readily done at the end of a call rather than during, simply because then the AI would have full access to the entire chart as iit is basically completed. This would also look at how complete a particular chart is and allow a much cleaner presentation to the billing side of the EMS world.
I can also tell from RocketMedic's post above that the post content was generated via voice dictation as my user ID isn't exactly conducive to transcription by most software.