Furthermore, falling is not binary, so there’s the question about where the boundary is. In an emergency situation, I don’t want to have to deal with any of this. It’s terrible UX.
"Call emergency services?"
"Yes" "No"
Although it removes the context of the watch' reasoning.
Reminder that the screen is a false positive protection mechanism and will only ever be seen in a specific context.
You have to solve for the lowest common denominator, not the hacker's technically correct preference.
This ain't about the watch falling, it's about someone possibly needing to call emergency services and contacts when they're hurt. "I'm OK" is a great prompt in that context, even if it's not an explicitly accurate response to whether the watch or wearer fell (which in turn is why the focus is not on the fall but on their need for help).
You did say "likely" but to provide an example: skiers who know they fell cannot comprehend the message trivially (not having noticed it's existence) and summoning help is often not the proper action to take. Call centers serving ski towns are swamped with false positives due to this issue.
“Why is my watching asking me to call 911 to get it repaired? My hip is broken!”
I can’t tell if you’re trolling me or not but I’ll assume you’re not.
If you didn’t fall, then you know it’s wrong.
If they are not laying in the floor then they already know it’s wrong.
"Fall detected. Call 911?"
“Huh? No, I didn’t fall. Don’t call 911”
Or
“I did fall, but I’m fine. Don’t call 911”
Apple's version is better. It's more accurate of what the feature does (guessing based on factors), and it's more human.
When a fall was detected the sample played: "A fall has been detected" When the caretaker was alerted: "The caretaker has been alerted" When the caretaker acknowledged: "The caretaker is on their way" ..
I dunno why its so hard. Talk to the user, re-assure them that the app is doing its job.