Apple seeks 911 dispatcher feedback over Emergency SOS skier misfires

Crash Detection

Apple is starting to collect information from dispatchers working in 911 call centers about incorrect calls from skiers' iPhones triggered by Emergency SOS and Crash Detection.

The introduction of the Crash Detection feature for the iPhone 14 and Apple Watch has led to a rise in calls to emergency services, with false positives leading to unnecessary calls and pulling attention from real emergencies.

In December, this became a problem for skiers, with reports that automated crash notifications from devices not actually involved in an emergency still placed calls. Without an answer from skiers when dispatchers attempted to speak to them, ski patrollers had to be dispatched to the location just in case.

Apple is now looking into the situation, according to the New York Post, with a spokesperson confirming the iPhone maker is talking to 911 call centers seeing a spike in automated 911 calls due to the introduction of Crash Detection.

While adding that feedback is being requested by Apple, the spokesperson declined to offer how the feature could be tweaked to reduce the instances of a falsely-detected car accident.

The report reveals the impact on call centers is fairly high in areas covering ski destinations. New York's Greene County 911 center saw a 22T% spike in hang-ups, open lines, and misdialed 911 calls in December 2022 versus 2021.

Jim DiPerna, 911 Communications Director for the county, reckoned there was a 15 to 25% year-on-year increase in calls "that very well could be generated by these Apple-generated and automated crash notifications."

In Pennsylvania's Carbon County Communications Center, there are up to 20 automated crash detection calls a day from local ski areas, which is described by Assistant 911 Manager Justin Markell as "taxing" for a team that is "already busy enough."

Follow AppleInsider on Google News

11 Comments

DAalseth

said about 1 year ago

That’s how engineering works. Design something, see how it performs in the real world, improve it based on what you learned.

the spokesperson declined to offer how the feature could be tweaked

Well, duh. They haven’t gotten the data yet. No point on speculating on the fix until you know what the problem is.

hexclock

said about 1 year ago

Leave your phone in a locker until you learn how not to fall. Problem solved.

dewme

said about 1 year ago

DAalseth said:

That’s how engineering works. Design something, see how it performs in the real world, improve it based on what you learned.

the spokesperson declined to offer how the feature could be tweaked

Well, duh. They haven’t gotten the data yet. No point on speculating on the fix until you know what the problem is.

What you’re describing is part of a proof of concept and evaluation phase of engineering. It looks like Apple’s engineers were somewhat lacking in their evaluation and consideration for what else could trigger false positives. Their apparent narrow vision very much reminds me of a scene in Monty Python and the Holy Grail movie concerning how to determine if a woman is a witch. Someone suggests they can reach this determination based on whether or not she floats in water. But unlike Apple, someone on their witch determination team asks the question:

“What also floats in water?”

Apple obviously came up with a set of criteria and algorithms that they assumed and verified would work for determination of their “witch,” i.e., occurrence of a crash. But perhaps they fell a bit short in probing the “What also floats in water?” part. Had they probed a little deeper they could have found it necessary to provide further safeguards, processes, or refinement to help reduce the likelihood of an incorrect determination. Or maybe they did fully understood the lack of precision and decided that the human benefits of a less accurate determination far outweighed the costs in time, money, and disruption of limited first responder resources, so they went to market with what they believed was their best solution for their highest level concern - saving people’s lives.

A measured, pragmatic, and biased towards saving lives approach is distinctly different than simply putting something out there and hoping for the best and knowing that feedback from its performance in the field will allow you to refine it over time. I don’t think Apple ever intends for its customers to serve as crash test dummies, even though customers do end up being testers/crash test dummies to some degree anyway because Apple, like everyone, doesn’t know what they don’t know. The line of distinction really comes down to Apple’s intentions and motivations, which are cultural influences over the engineering and product development process but not part of the engineering process itself.

In my opinion, none of the false positives identified thus far in Apple’s deployed crash detection performance seem like radical departures from scenarios that would or should have been considered if Apple probed a tiny bit deeper into the “What also floats in water?” line of inquiry.

mikethemartian

said about 1 year ago

Doesn’t the OnStar system connect a person to an OnStar dispatcher first to determine if an actual emergency exists before calling 911?

mikethemartian

said about 1 year ago

DAalseth said:

That’s how engineering works. Design something, see how it performs in the real world, improve it based on what you learned.

the spokesperson declined to offer how the feature could be tweaked

Well, duh. They haven’t gotten the data yet. No point on speculating on the fix until you know what the problem is.

Actually in engineering you usually have very strict requirements on how a device will perform under different conditions and you test accordingly both during the design verification phase and the production phase.