Apple Executives Offer Closer Look at iPhone 14's Crash Detection

[Allthings-I]

So many negative comments out here, I’d rather know my loved ones are found/provided care asap should such an emergency occur while driving out there, including to myself.

Most of the people commenting probably haven’t been in a bad accident, an accident doesn’t have to be as bad to cause serious injury where a person may not be able to move.

Now Apple has some work to do and I’m sure crash detection will improve with time as more real world data is provided to Apple and sensors in future generations improve.

For me I’d rather have then not have this feature and an underestimated feature for those that think its useless.

 

[caribbeanblue]

Lorem ipsum dolor sit amet, consectetur adipiscing elit.

 

[theluggage]

So many negative comments out here, I’d rather know my loved ones are found/provided care asap should such an emergency occur while driving out there, including to myself.

Sorry. The emergency services were late to your accident because they were dealing with a string of bogus calls caused by an "oversight" in iOS 18's crash detection.

The emotive "well, it may save a few lives so we should do it" argument is terribly dangerous when it's used to reject criticisms as negative. With any feature like this, false positives have serious consequences which affect others - especially with the rate at which new iDevices get adopted. Emergency services are rarely over-staffed or over-funded and if people's brand new iPhones suddenly start phoning in non-existent car crashes it is likely to exceed the services' capacity to respond, which can easily cost lives in itself.

I don't know how much effort Apple put into exploring the unintended consequences of this feature, calculating the false positive rate & its consequences, doing scientific (not anecdotal/emotional) risk/benefit analyses and working with emergency services to prepare them for the rollout - but the "roller coaster" report suggests that they didn't sit down and have a "what could possibly go wrong?" brainstorm.

 

[Expos of 1969]

The thing is it that useful? Are there really statistically significant numbers of car accidents where the driver/passengers are incapacitated, there is no witness or other help, and there’s also phone signal?

It seems to me this has come primarily from a marketing standpoint. They knew that safety features such as fall detection were good for sales of the Watch and pushed for something safety related for the iPhone. And crash detection fit the bill.

Just look how heavily it was marketed at the keynote with mocked up crashes etc. This could have been a small additional feature but it was centre stage. Buy the new iPhone to save your life.

In reality it seems to be causing more problems that it solves.

Somewhat similar to the "But we must save the children" CSAM nonsense.

 

[notabadname]

False alarms happen across the spectrum of “detection” and alerts, but we also don’t recommend disabling home security systems, fire alarms, severe weather alerting systems. I was at a train crossing where the crossing gates malfunctioned and lowered and stopped traffic for a non-existent train. They have not removed the crossing guard as a result. When the false alarms are statistically low - as in this rollercoaster alert, the life-saving rewards are well worth it. I happily accept the statistically low rate of false alarms if it saves one human’s life.

 

[dk001]

After seeing this in vids, in podcasts, and threads like this, here’s a thought: Where is the ROI? Apple and other companies don’t do this stuff out of the good ness of their money grubbing hearts.

How many lives will this really save?
At what cost?
I can hear the “if it saves just one…” … I get that however that is not the going in premise for any corp. It all comes down to $$$

Answer: Sales (Markewting)

 

[compwiz1202]

Engineers have a hard time imagining how customer's can misuse the products they design because no sane engineer would do that themselves.

For most of the games I betaed, even with big beta groups, it barely took the public any time to find more bugs/exploits.

 

[compwiz1202]

After seeing this in vids, in podcasts, and threads like this, here’s a thought: Where is the ROI? Apple and other companies don’t do this stuff out of the good ness of their money grubbing hearts.

How many lives will this really save?
At what cost?
I can hear the “if it saves just one…” … I get that however that is not the going in premise for any corp. It all comes down to $$$

Answer: Sales (Markewting)

Lives are priceless so $$$ are irrelevent. The issue is for how many it saves, how many are lost to false positives taking resources from a true emergency?

 

[I7guy]

After seeing this in vids, in podcasts, and threads like this, here’s a thought: Where is the ROI? Apple and other companies don’t do this stuff out of the good ness of their money grubbing hearts.

How many lives will this really save?
At what cost?
I can hear the “if it saves just one…” … I get that however that is not the going in premise for any corp. It all comes down to $$$

Answer: Sales (Markewting)

Do a google search on the trolley problem.

While you are certainly within your right to be cynical, there may be another less cynical viewpoint. And that is that this technology of detecting motion and stops will be deployed in other products for other purposes. Because it’s good for the customer. If it’s good for the customer it’s good for Apple and that’s how it translates into $$$.

 

[jweinraub]

A couple thoughts.

1. False positives causing erroneous calls to emergency services are a cost. Notifying emergency services when an actual crash happens is a benefit. If false positives are relatively rare, then maybe that's an acceptable cost. But the feature is so new that we don't yet know. But Apple will accumulate more data every month on the frequency of correct vs. false positives so it can better assess the actual usefulness of the feature. (Hopefully it will share this information with us.)

2. Apple could make the crash-detection feature "opt-in," and require anyone who opts in to first read some splash screens about the importance of disabling the feature when you visit amusement parks, and how to disable the feature. Of course, some people will forget to do this, so...

3. Apple could automatically disable the feature while a user is in an amusement park. Apple knows where all the amusement parks are (because Apple Maps) and location services are required to be enabled to use crash detection, so this seems an easy way to eliminate many false positives.

Unless of course you have a collision in the car park? Or on the motorway adjacent? Geofencing is still not acceptable because people will expect it to work there when it doesn't. I will use reminders to alert me when I get home and often it triggers 300 m from my home.

 

[bizack]

For most of the games I betaed, even with big beta groups, it barely took the public any time to find more bugs/exploits.

ML is not a process of Engineers, PMs or Designers imagining use cases and then beta testing to find bugs. The goal is to model the environment of a moving vehicle before, during and after a crash. The use case is irrelevant because there are near infinite possibilities of what a car crash looks like. The testing is to improve the model and to improve usability.
Games have predetermined, expected behavior. Although a game can have many different outcomes, the environment is created and therefore constrained by the game developers and producers.

 

[compwiz1202]

Do a google search on the trolley problem.

While you are certainly within your right to be cynical, there may be another less cynical viewpoint. And that is that this technology of detecting motion and stops will be deployed in other products for other purposes. Because it’s good for the customer. If it’s good for the customer it’s good for Apple and that’s how it translates into $$$.

Love trolley problem. I was going to say that not having it definitely kills the ones that go unconscious with an injury that needs immediate attention with no one else around. But false alarms may or may not kill someone in that situation delayed by the false alarm

 

[dk001]

Lives are priceless so $$$ are irrelevent. The issue is for how many it saves, how many are lost to false positives taking resources from a true emergency?

Missing the point though I do agree.
Why would Apple create this feature? There has to be an ROI associated with this before they sink millions into it.

 

[dk001]

Do a google search on the trolley problem.

While you are certainly within your right to be cynical, there may be another less cynical viewpoint. And that is that this technology of detecting motion and stops will be deployed in other products for other purposes. Because it’s good for the customer. If it’s good for the customer it’s good for Apple and that’s how it translates into $$$.

Hence my question.
When a new feature like this is proposed, there is an associated cost benefit that has to be on the positive side.
Not cynical, just curiosity. CBA is always a part of things like this.

 

[theluggage]

Love trolley problem.

The trolly problem is all about introducing different schools of philosophical thought, and is so abstracted that it has little to do with solving any sort of practical problem. The whole point seems to be that there is no "right" generalised answer and as soon as someone gives an answer you shift the goalposts by changing the context - all good harmless fun to introduce utilitarianism in philosophy 101 but its really, really worrying when the ML people start talking seriously about it in the context of self driving cars etc...

Crucially it says nothing about the sort of risk assessment and probabilities that happen in the real world: you're gifted with perfect foresight: pull the lever 1 person will definitely die, do nothing 5 people will definitely die - everything is literally on rails (and not by accident - there's a reason it's not the "donkey cart problem") - no pesky uncertainties or doubt and certainly no concept of false positives (how do you know the trolley is out of control?)

(Hint: #1 keep the car/trolley under control - if it turns into a hurtling mess of twisted metal, all bets are off as to who it is going to hit!)

 

[RodThePlod]

Missing the point though I do agree.
Why would Apple create this feature? There has to be an ROI associated with this before they sink millions into it.

Apple are equipping their phones and devices with all sorts of sensors that, together, will help to enable the AR/VR experiences that they're developing. They're definitely sinking millions into those initiatives - so maybe crash detection is just a useful feature/side benefit that's enabled by those sensors? Just a random thought!

 

[bizack]

Apple are equipping their phones and devices with all sorts of sensors that, together, will help to enable the AR/VR experiences that they're developing. They're definitely sinking millions into those initiatives - so maybe crash detection is just a useful feature/side benefit that's enabled by those sensors? Just a random thought!

Apple spends over $20B/year in R&D. They're pouring billions, not millions into sensor-based products and initiatives.

 

[I7guy]

Hence my question.
When a new feature like this is proposed, there is an associated cost benefit that has to be on the positive side.
Not cynical, just curiosity. CBA is always a part of things like this.

It’s the positive side of CBA is saving a life. Same as fall detection and afib detection and ekg functionality.

But let’s be honest, unless one of us is in the know we will never know the rational behind where and how these features get chosen.

To me, it’s part of a comprehensive ecosystem which adds value to the consumer. Adding consumer value increases revenue.

 

[McMountaineer]

Maybe for future it will be a good idea to add an airbag to the iphone to prevent damages when falling from a table or your hands accidentally.

 

[I7guy]

Missing the point though I do agree.
Why would Apple create this feature? There has to be an ROI associated with this before they sink millions into it.

Occams’ Razor might help here. Your inability to discern the ROI associated with this feature is not relevant. The fact that apple implements features like this makes me happy to be a shareholder, customer and observer.

 

[Mousse]

False alarms happen across the spectrum of “detection” and alerts, but we also don’t recommend disabling home security systems, fire alarms, severe weather alerting systems.

Those systems have been debugged to the point where false alarms are few and far between. Apple's crash detection hasn't. False positives on a rollercoaster, no crash detected in a junkyard crash test. Apple's crash detection ain't ready for prime time; its usefulness questionable at best. When they can get the number of false positives down to a reasonable level, then they can tout its usefulness. One more thing, Apple's system is redundant. When a crash occurs, there are usually witnesses and a lot of cars have crash detection built in (ie. OnStar).

 

[notabadname]

Those systems have been debugged to the point where false alarms are few and far between. Apple's crash detection hasn't. False positives on a rollercoaster, no crash detected in a junkyard crash test. Apple's crash detection ain't ready for prime time; its usefulness questionable at best. When they can get the number of false positives down to a reasonable level, then they can tout its usefulness. One more thing, Apple's system is redundant. When a crash occurs, there are usually witnesses and a lot of cars have crash detection built in (ie. OnStar).

Debugged? False alarms for security systems happen regularly, from kids activating, to the dog or a storm. The Apple detection, like OnStar, is intended for when you are not near a “witness”. And OnStar needs a subscription And have their own disclaimers. OnStar also mentions vehicle damage and an operating electrical system, as in the Synch’s.

OnStar services. The most basic plan is the Connected Vehicle plan and this costs $24.99 per month. With this plan, you get services like vehicle location and remote personalization. The next plan is the Safety and Security plan, which costs $29.99 per month. This plan includes OnStar's safety features, like roadside assistance and automatic crash detection but does not include the connected services found in the Connected Vehicle plan.
To combine the features of the Connected Vehicle plan and the Safety and Security plan, you'll want the Essentials plan. This plan costs $39.99 per month and includes all of the features found in the other two plans

For Ford Synch’s crash detection: The vehicle's electrical system (including the battery), the wireless service provider's signal, and a connected mobile phone must all be available and operating for 911 Assist to function properly. These systems may become damaged in a crash. The paired mobile phone must be connected to SYNC, and the 911 Assist feature enabled, in order for 911 to be dialed. When the feature is ON, 911 Assist uses your paired and connected mobile phone to assist.

But certainly, feel free to disable yours. 👍🏻

 

[dk001]

Occams’ Razor might help here. Your inability to discern the ROI associated with this feature is not relevant. The fact that apple implements features like this makes me happy to be a shareholder, customer and observer.

Assuming much? lol I'm not.
It's an honest question. I do projects and product dev and this is always a question.

 

[I7guy]

Assuming much? lol I'm not.
It's an honest question. I do projects and product dev and this is always a question.

You can’t ask an honest question that only apple can answer. If I misread your post and it was all about speculation as to why and roi then I apologize. But if you were asking what apples rational and roi and cba no one save few in the know can answer that. And if you a seasoned veteran can’t discern cost benefit analysis or roi, how are us mere mortals supposed to?

If you’re a shareholder wanting to ensure apples feet are held to the fire you should bring it up at the next meeting.

This question would be better asked jn the apple tech forum.

 

[Mousse]

Debugged? False alarms for security systems happen regularly, from kids activating, to the dog or a storm.

Those are forces outside the system. Apple's flaws lies within the faulty software. They can brag about its usefulness in a few years when the work out all the kinks.

I guess some people think Apple's system is good enough the way it is. I just hope Apple engineers don't have the same attitude as Apple zealots. If Apple only listens to the Yes Man zealots, Apple Maps would still be directing fanbois onto nonexistent streets.😒

But certainly, feel free to disable yours. 👍🏻

Will do. I might turn it back on once it works as advertized. My wife uses Apple Maps as her exclusive GPS now, something she would never have done a few years back.