Ask HN: Why is there no way to lower amber alert volume on headphones?
I want to contribute to community efforts at finding missing kids and be ready for local disaster situations. But as someone who wears headphones my ears are often rattled for hours by these alerts.
This can't be an isolated issue as google indicates multiple lawsuits about this. https://9to5mac.com/2022/05/17/apple-hit-with-lawsuit-over-ear-shattering-amber-alert-volume-through-airpods-pro/
The link is for apple, but as an android user I face the same issue. What I don't understand is why
emergency notifications don't respect media volume? Can someone explain the technical logistical reasons for this?
153 comments
[ 2.6 ms ] story [ 213 ms ] threadTurn off amber alerts; they're entirely useless and you can now turn them off without turning off the tornado warnings (almost entirely useless, too).
There may be a way to turn off the "Watch" alerts and still keep the "we are gonna die" alerts. The one time we got the green sky of doom the phone did explode in a way not normally seen, so I assume they know something about crying wolf.
As for amber alerts being useless - something like 90% (?) of the time it's a custody dispute where one parent hasn't returned the kid at the appointed time.
https://www.weather.gov/nwr/eventcodes
Once you know about them, they're relatively easy to spot (look in areas near power plants, downstream of dams, etc).
You can find some here and use them as ringtones!
https://www.whelenmassnotification.com/recorded-message-libr...
The government mandated stuff just needs to be turned off because it's broken.
In Texas my iPhone has this alert granularity: AMBER Emergency Public Safety Test
If the local government is sane, they send WARNING and (unnamed) only - (unnamed) being the YOU ARE ALREADY DEAD MAKE YOUR TIME notices.
Then if you want WATCH you set your weather app to notify you on those.
I can’t speak to tornado alerts, but the Amber Alert child alerts were becoming excessive and useless in my state.
I was getting alerts from many hundreds of miles away, often while trying to sleep. Other times they’d forget to put relevant information in the first alert so they’d send several more to follow up. When we finally got a series of multiple Amber alerts in the middle of the night, most of the people I know turned them off.
The system may have started with good intentions, but it was mismanaged and abused so much that it became useless.
It should be possible to improve policies without introducing dependencies to crappy JavaScript frameworks.
To be honest I wonder if the people in charge were starting to get a kick at waking up so many people, or if they were engaging in some sort of malicious compliance, e.g., maybe the people responsible for sending the alerts out are also pretty pissed off at being woken up to send them out. The stated reason for them existing leaves no room for there to be any reason to be sending these out at 3am, if neither the abductee nor the putative abductor were known to be nocturnal.
And then they don't even send another alert saying she was found safe and sound sleeping at her dad's house.
A missing child in a custody dispute alone is rarely enough to trigger an alert.
One of the leading types of child abductions leading to alerts is vehicle theft with a child in the vehicle. Not a custody dispute.
"The vast majority of child abductions are carried out by a relative or close acquaintance of the victim, often a divorced parent who was not granted custody... about 20 percent involved a kidnapping by a stranger or slight acquaintance of the child. In the other 80 percent of cases, the youngsters were taken by a relative (most often a parent) or an acquaintance (frequently a babysitter)."
source: https://psmag.com/social-justice/amber-alerts-largely-ineffe...
"Presumably because of that need "to make an intuitive, rapid decision," as Griffin puts it, AMBER Alerts often turn out to be false alarms. Of the alerts the UNR team tracked, about 20 percent involved a kidnapping by a stranger or slight acquaintance of the child.
In the other 80 percent of cases, the youngsters were taken by a relative (most often a parent) or an acquaintance (frequently a babysitter). While such incidents can be traumatic to both the child and the custodial parent, they are routinely resolved peacefully."
I doubt it. The messages were just incremental additions of information as it was added: First message had something about the area. Second message had the color of the vehicle. Third one added another detail or a correction or something, I can’t remember.
I just remember that everyone in my local office was visibly tired the next day and we all shared the information to turn the alerts off.
Maybe go with the UX best practice of allowing the user to choose how they want to be alerted? Is that so hard? Government can't get basic shit like this right, I wonder how they're handling the rest of it...
Isn't part of the issue that when a child is taken, the assumption is that the kidnapper will immediately flee the area? I think hundreds, plural, might be excessive, but it wouldn't be unreasonable to notify people within a certain distance based on high speed and how much time has passed; maybe 50-100 miles?
I can see getting annoyed at multiple messages, but it's also important to remember that working with time sensitive emergency services is a extraordinarily difficult thing. It's like, everything is important all at once, and your mind can only process it all so fast. It's extremely time sensitive, and if you mess up any of the details, you are directly delaying the proceedings.
I haven't had to issue anything like an Amber Alert, thank God, but I did have to relay some information from a victims family to a supporting organization, and basically screwed it up the first time around. I ended up having to end the call with the family, call the organization to see what they needed, call the family back, then call them again with the relevant details. It was an incredibly brutal thing. You can feel the rawness of the emotion through the phone for the people. In my own experience, like, I could hear people crying and screaming in the background. It's this insanely personal moment for them, but you kind of have to force your way in the middle of it to make something happen. The person who I started off with had to hand the phone over because they thought they were about to pass out. It's really a very difficult to communicate in the midst of a crisis.
Anyway, I'm not sure which area you're in, but from my own experience, the very few times I've gotten one, it was always very succinct and clear. Child's name, description of their clothes, make/model/license plate of car, which direction they were headed, and when it happened.
I certainly don't blame you for turning off the alerts, or for being annoyed at the emergency services. The point wasn't to finger wag or whatever, it was just to say that it's extremely difficult to do, let alone do well.
States vary as to how effectively they run their alerting systems, and how well they follow best practices, but that is ultimately an issue for the voters.
In my own assessment, I classify that person to be dangerous. A stranger takes a child for an unknown reason. A parent takes a child, willingly harming the child physically or mentally in hopes of doing harm to a person they hate. I would assume they are willing to do far more than mere mental harm to the child, if it means "winning" the fight. I don't have any data to back that up, and it's entirely possible I'm wrong. I just don't see an Amber Alert being raised because a parent took the child to be any less serious than if a stranger took the child.
I would imagine that some number of Amber Alerts are a kind of "false positive", something like failing to drop a child off on time (out of gas, severe traffic, etc.) and their phone dying, turning into a big ordeal, but I just wouldn't discount the situation just because the suspect is a parent. Maybe that shouldn't meet the threshold, but I just personally don't see it as being a big enough nuisance to be bothered by it. Again, it is easy for me to say that though, since I don't get those sorts of alerts very often, maybe 2 to 3 times per year, if that.
I do also completely agree though, that as it relates to headphone usage it should be changed. I think that if you are listening on your headphones, it shouldn't come through your headphones at all. Just play the noise on the phone speaker - it'll probably still draw the users attention. I might be asking too much though, as there was that 911 bug [1][2] a while back.
[1] https://news.ycombinator.com/item?id=29492884
[2] https://old.reddit.com/r/GooglePixel/comments/r4xz1f/pixel_p...
Sure, that might not work well when I'm on travel, but when I'm on travel I'm typically in places with a ton of built-in mass notifications systems (airports, metro & train stations, roadside state DOT banners) so I don't feel I'm missing out.
https://nymag.com/intelligencer/2013/07/ambert-alert-phone-4...
Even older parents called their kids to find out how to make sure it does not happen again.
Many abductions are vehicle thefts where the thief takes off without noticing or caring that there is a child in the car.
In one recent case, the thief threw the baby out the window into a ditch. Citizens who received the Amber alert found the child.
It IS fun to get the weather warning and then run outside and hear the air-raid siren begin to warm up.
Something similar could easily be done.
I wonder if the "limit max volume" setting elsewhere does anything at all. I suppose it's hard to test.
I CANNOT have my phone making noise, so it has to be fully off. If there was a choice for silent notifications, I would choose that.
What we need is political pressure for a pay-to-play solution. You fire an alert at my phone, that's $100 deposited within a week against my phone bill or offered via mail for the pay-as-you-go burner phone types.
Then if the phone goes off I know its something real. Right now it's just a waste of resources and annoyance, probably killing more people than it is saving via sleepy or distracted drivers or medical professionals.
But on the other hand it's possible to turn them off so maybe not.
Maybe you could tweak something here?:
https://cs.android.com/android/platform/superproject/+/maste...
(although it looks like there's some code that already intends to do this function)
We don't have amber alerts in my country (or at least I never got one). However Mi wireless earbuds + Nokia X10 for some reason reset to a painfully loud 100% volume when receiving a call. I can lower the volume in settings, but on next call it resets to 100% again. Same phone with a different Bluetooth headphones and same headphones with different phone work fine.
I don't know the whole software stack to point fingers whose fault it is, but there must be a bug either in this specific phone's Bluetooth implementation, Android 12 or the Mi earbuds.
Emergency volumes are supposed to be loud, and override the volume settings.
The thing making that override doesn’t know what the actual audio output channel is, or what “loud” means on that channel.
Users “reverse opt in” to amber alerts by not turning them off. Most people just these off. There’s way too many of them, and the vast majority of these are a custody dispute escalation. Don’t feel bad for shutting these off.
With this fact pattern the issue never makes it to the top of the stack to solve.
I've never encountered this before, and I don't see any option for it on my phone.
I'm an Android user in the UK, if that helps.
There's also silver alerts, where you do the same for a missing old person.
A plausible thing to say, but do you have any numbers to back this up?
https://en.wikipedia.org/wiki/Amber_alert
[0] https://en.wikipedia.org/wiki/Amber_alert#Namesake
They amusingly alert when they're not supposed to.
https://en.wikipedia.org/wiki/2018_Hawaii_false_missile_aler...
It’s deployed in a lot of countries.
Search this thread for “Ontario” (Canada) for example.
A Silver alert is very similar and is sent out when an elderly person (esp. with memory issues) is missing.
(A divorced parent actually kidnapped their own kid, or is late dropping a kid off.). Less commonly, it will involve other family members. (A parent is a drug addict, so the grandparents/uncle take the kid.)
In a vanishingly rare subset of the alerts, it is a kidnapping where a kid is taken by a stranger.
The phone settings are usually called "emergency alerts" or similar.
Amber alerts are bad. We get them now and them in Ontario at 3am, sent to the entire (rather large) province. This is idiotic. The probability that someone would be awake and in a position to observe an abducted child at 3am is negligible. The probability that someone woken at 3am will not be able to get back to sleep, drive to work in the morning when drowsy, and consequently get into an accident that kills someone, is not negligible.
Such alerts should be confined to warnings that people really do need to pay immediate attention to. Like a train accident that has released toxic gas, requiring immediate evacuation, even if it's 3am. As it is, an alert at 3am will be ignored, since the system has been abused by the idiots running it.
[1] https://cops.usdoj.gov/BlueAlert
Yes, they absolutely are. Just a couple weeks ago, local law enforcement sent out an incredibly vague shelter-in-place active-shooter alert across the entire Twin Cities area for a situation in a suburb that was resolved about five minutes later[1]. Schools were locked down, buildings closed, etc. And wouldn't you know it, there's an election in two weeks where the Republicans, who the police unanimously endorse and campaign for, are running on a "crime is out of control" narrative! Crazy coincidence, that.
[1] The entirety of the message was, "Shelter in place - Homicide suspect at large described as 17 year old white male." No location specified, sent to the entire urban region. https://www.startribune.com/shelter-in-place-alert-about-hom...
This was completely useless and actually distracting considering that I was in heavy traffic at the time.
An honest mistake is an accident that isn't repeated, or if repeated, happens in a random manner rather than exhibiting any particular pattern (seems to express neither a conservative nor a liberal worldview, nor any other potential agenda, consistently).
If someone has a desire to do something like use some position or tool they happen to have access to for some purpose of their own rather than it's intended use, and lacks the integrity to avoid doing so on their own, that desire and that lack of integrity doesn't change or stop even if they get chastised for some misdeed. They still want to do things like that, and they still don't see what the problem is, and they still do the same things, merely trying to do a better job of camoflage and deniability.
Some states will not send alerts overnight and others will; it is a matter of state policy - and ultimately an issue for the ballot box.
Look how parents who don't handcuff their children to their wrists so they can keep an eye on them 24/7 are often treated.
The US system has different levels of alerts: Presidential/National Alerts, AMBER Alerts, Extreme Weather Alerts, Dangerous Weather Alerts, etc. On my Pixel 6, everything but the Presidential/National Alerts can be disabled. Same with my wife's Galaxy S22. I assume the same for iPhones.
However the Canadian system (which is heavily based on the American system) is setup differently. All alerts are sent as 'National/Presidential Alerts'. None of the other alert levels are used.
https://amberalert.opp.ca/
The Canadian is based on the US system with different levels: National/Presidential Alert, AMBER Alert, Extreme Weather Alert, Dangerous Weather Alert, etc.
However in our system, all alerts are sent as 'National/Presidential Alert'.
We are on Telus/Videotron.
That's the dumbest thing ever. A presidential alert basically means "nukes incoming."
That's kind of understandable, though. If the nukes really are incoming, you want to make sure your alert system actually works, and the only way to really do that is to test it.
[1] few disasters, man made or natural, reach national scope. Even a grid failure would likely only reach half the country, but do you need a national alert for that? What useful content could it have?
Many inbound nukes is about all I could see it used for, but I expect that would cause more panic and be less useful than not sending a message... And I'm not convinced a decision to send a message would be made in a reasonable amount of time. State or local messaging seems more likely.
What now? Sure it would cause less panic--in same the way shooting someone in the head unannounced would cause less panic than giving them warning to duck. If the nukes are incoming, "avoiding panic" is the wrong priority.
> And I'm not convinced a decision to send a message would be made in a reasonable amount of time. State or local messaging seems more likely.
IIRC, they're triggered directly for NORAD (or whatever it's called nowadays). There was actually a false alarm alert in the 70s, and it did get out quickly (google "code word hatefulness").
Well it got out quickly, but the cancellation was slow.
I have no doubt that the alerts would go out quickly, if there was a decision to send them (and assuming the means to send them is still around). More and more of these systems has gotten automated so there's no one at the local stations who would need to decide to send or not. I'm just not sure that the 25 minutes or so between detection and impact is enough time to decide to send an alert, amidst all the other activities that would certainly need to be done.
That's why I disable them totally. In my experience, they're always just a description of a car, and realistically there's literally nothing I can do with that information. When I'm at home, I'm not going to check license plates, and when I'm driving I'm not going to check my phone (nor remember a random license plate number for more than 10 seconds to check anything).
They're also always "non-custodial parent took the kids," and I'm pretty sure the police in that case can quickly discover much better information than I could ever give them. There was one recently in my area that a friend of mine received. Later I read in the paper that they just tracked the guy's cell phone to find him.
The USA has 332 million people, at least 3 of which are children, and gets about 200 Amber alerts/year (usually not nationwide, only locally, which can be an entire state).
The numbers line up if you count the "app notice" only ones.
> An Amber Alert (or child abduction emergency alert) is a message distributed by a child abduction alert system to ask the public for help in finding abducted children. The system originated in the United States. [1]
Alerts are distributed via "pagers, faxes, emails, and cell phones with the information immediately posted on the Internet for the general public to view" and have the aim of "ask(ing) the public for help in finding abducted children".
A report indicates around 200 Amber Alerts are issued per year in the US: https://www.missingkids.org/content/dam/missingkids/pdfs/amb...
[1] https://en.wikipedia.org/wiki/Amber_alert
FYI, this isn’t unique to the USA. Canada has serious problems, too, especially in Ontario. There are also various references to Belgium and other countries in this thread.
Didn't even realize you could get them on a flight. Must be because of the in-flight wifi.