I believe I touched on that in my last sentence, but I can elaborate:
That will probably happen – neural networks are approximation algorithms. It’s a question of how often that happens.
What percentage of the calls get misclassified? And what’s the threshold percentage that would be needed for the triage bot to be a net positive?
It sounds like they have an idea of these numbers based on data collected from the non-emergency line, so it’s not like they’re just blindly jumping into this.
EDIT: I just realized that I did not actually answer your question of “what happens”…
I imagine the caller would just interrupt the AI’s answer (e.g. “No not that,” “HELP,” “Give me a human,” “FUCK!” etc.)? That seems like the natural thing to do.
To be clear: I don’t know anyone at Carbyne or OPCD. I can only offer speculation.
I don’t think there’s any net positive that would account for not answering an emergency call at all.
Exactly. That’s what they want to solve.
TL;DR: Even when callers reach the triage bot, they can still reach a human much faster than without the triage bot.
Comparing again:
WITHOUT the triage tech:
NOBODY (or nothing) answers the call for a long while, because the caller is stuck in a very long queue of calls waiting to tell them about the same emergency
WITH the triage tech:
AI bot answers the call instantly and probably knows how to help because the call is probably about the same emergency that 95% of the other calls are about
so 95% less spam for the human operators to get through
If the call is NOT about the same thing as the others, the caller can simply say that (i.e., “no”), and they reach a human within, say, 5-10 seconds because the operators aren’t busy trying to get through the spam calls
They chose this tech because it has already proven to be a net positive on their non-emergency line.
Do you honestly think that nobody has thought about that and solved that problem already? Even with all the engineers involved, and after all the real-world testing, you’re the first to have considered that scenario?
Even consumer-grade products like ChatGPT can be interrupted while talking. We’re talking about an AI implementation, not some rigid set of if...else statements.
You talk as if they don’t already have customer facing implementations in effect, and that those all suck and don’t work worth shit. You also seem to think the “engineers” involved are custom doing anything and not just some sales person slapping a half baked product on every problem. When you get paid per deployed program, everything looks like a problem to solve with said program.
This is an issue of not having enough 911 dispatchers, due to the unwillingness to pay for them. The idea of putting in a chatbot, that can not even speed up diction let alone have any empathy is just wildly inappropriate in this situation. The fact is that these LLMs will (like in current deployments) most of the time have to pass the call to a person, drastically increasing time on the phone before action is taken. This is already seen in almost all LLM supported call centers, but a 911 dispatch is not a telecom company and more time on the phone is more death, injury and suffering.
Do you honestly think that nobody has thought about that and solved that problem already? Even with all the engineers involved, and after all the real-world testing,
That is the thought process of a child. If your defense of something with obvious issues is “Well I’m sure they know what they’re doing, because if they didn’t it would be bad”, I have some really, really bad news for you about the world.
If this was thought through by a capable adult, it would’t be implemented at all. The very fact that this solution is on the table is proof of the breakdown of rational thought. Or more likely, greased palms.
Your blind appeal to authority would hold more water if we haven’t already seen national adoption of AI products in emergency services that that have catastrophically failed to perform their basic functions, leading to real world harm. Audits of Flock cameras deployed in several cities found an error rate ranging from 33-70%. People are being arrested for things they didn’t do because the AI doesn’t work and nobody cares. People are being stalked because nobody thought of basic security controls.
We’re talking about an AI implementation, not some rigid set of if...elsestatements.
Even worse, we’re talking about an AI implementation of if…else statements. Regular if else statements are consistent. As soon as you throw AI into the mix they become unpredictable. You introduce a failure mode that didn’t exist before that can and will have real consequences.
So… you do assume that you’re the first to think of this problem, or…?
If you assume that your brain is somehow superior[1] to everyone involved in the R&D, engineering, data analysts/QC, etc., then this may come as a surprise to you: You’re not even the first Lemmy commenter to think of it.
Literally multiple people here already thought about that and commented before you did. They’re not the first either; the author of the article herself even thought about it.
This idea is among the most obvious potential problems that come to mind within seconds or minutes of just hearing about the concept. It’s not a dumb idea by any means (it’s completely valid), it’s just not a particularly smart or special one.
911 is not the first department in that city[2] to use this tech, and that city isn’t even the first to adopt it. It’s silly to assume that such an easy-to-solve problem hasn’t been addressed by now, or that it somehow has not yet happened in other deployments.
To be clear: I make no assumptions that your brain is not superior to theirs either – I couldn’t possibly know either way. AFAIK, I’ve never even met you nor them… But I think it would be a safe bet that you don’t know either. ↩︎
BTW, have you heard cajun accents? There’s no way this bridge hasn’t already been crossed during their 311 trials. ↩︎
As someone who worked in the industry, your trust is wildly misplaced. I guarantee someone has already said internally this will not work and was shushed by a sales manager.
And as far as “I am sure they tested accents” goes? HA they likely noted it was a mess and moved on. You think the goal is a functioning system, but it is not it is just to make the sale.
Why does it have to be AI? Why can’t it be “press 1 for the highway car fire on 42, press 2 to be transferred to an operator”? Surely that’s much less likely to fail?
Ah, right, I suppose it’s contextual. You would also have to read the comment that it’s responding to. To summarize: it might not have to be AI.
However, 911 is a bit too dynamic for your solution to be viable as-is. For example, “Hey Siri/Google, I can’t move my hands. Call 911.”
Your idea is probably a decent starting point for brainstorming a solution — it just needs more thought, research, testing, etc. And I wouldn’t be surprised if a similar product already exists. (I’d actually be surprised if it doesn’t already exist.)
I believe the argument is that the probability of misclassifying an arbitrary segment of speech as the very particular answer “yes” is sufficiently low (with reasonably trained, current models) that we may ignore it under practical circumstances.
I find the non-zero probability of misclassifying in a little disquieting, but I suppose one way of looking at it is that this will probably increase the expected value of lives saved; the very probable case (calls are correctly filtered and a smaller proportion is passed through, allowing operators to respond to more new emergencies) may save a lot of lives, whereas the very improbable case (something that is not “yes” is misclassified as such) may endanger a few.
One thing that would worry me about just looking at expected values is the possibility of bias against a particular group of people or emergency type, but to me that seems unlikely in this case.
Edit: it also just occurred to me that if you are woefully unfortunate, you can probably just call again if you accept the assumption that there is a high probability the answer is yes given the model classified it as such. It might be more or less random chance?
Yeah I just don’t buy that there is a case where any “net gain” here is justified. These are people’s lives that you’re playing with by reducing them to what amounts to a balance sheet in the name of saving money.
The idea that you would create the possibility of denying someone emergency care that didn’t exist before, no matter how improbable (probably not even that improbable, given the propensity even the most advanced frontier AI models have for getting things wrong or hallucinating entirely despite simple instructions) just to save 10-15 seconds on average for other callers is not only absurd on its face, but morally bankrupt. You can personally ignore it because you don’t live there and it’s not the lives of you or your family at stake. I think those who this system could fail would not be able to ignore a flaw like that when it happens to them, and the idea that their peril was “highly improbable” will not be of much comfort to them.
The problem is that there are staffing shortages. The solution is hiring more staff. Trying to cheat our way out by implementing a system prone to unmitigatable flaws that could have life-altering or even life ending consequences isn’t a solution, it’s dystopian.
how long could the queue be? If someone can’t wait 60 seconds for the queue, but can take the 8 seconds to ask if it’s about X incident. I do imagine it will be smart enough to check their approximate location before giving the prompt
I believe I touched on that in my last sentence, but I can elaborate:
That will probably happen – neural networks are approximation algorithms. It’s a question of how often that happens.
What percentage of the calls get misclassified? And what’s the threshold percentage that would be needed for the triage bot to be a net positive?
It sounds like they have an idea of these numbers based on data collected from the non-emergency line, so it’s not like they’re just blindly jumping into this.
EDIT: I just realized that I did not actually answer your question of “what happens”…
I imagine the caller would just interrupt the AI’s answer (e.g. “No not that,” “HELP,” “Give me a human,” “FUCK!” etc.)? That seems like the natural thing to do.
To be clear: I don’t know anyone at Carbyne or OPCD. I can only offer speculation.
I don’t think there’s any net positive that would account for not answering an emergency call at all.
Can you really justify someone not being able to reach help at all in place of everyone being able to reach it albeit slower?
As a lifelong first responder, I couldn’t get behind something like this at all.
:sigh: Okay, I’ll try to break it down even more…
Exactly. That’s what they want to solve.
TL;DR: Even when callers reach the triage bot, they can still reach a human much faster than without the triage bot.
Comparing again:
WITHOUT the triage tech:
WITH the triage tech:
They chose this tech because it has already proven to be a net positive on their non-emergency line.
None of that applies if a caller just can’t get through at all because the bot mistook “no” for “yes”
Do you honestly think that nobody has thought about that and solved that problem already? Even with all the engineers involved, and after all the real-world testing, you’re the first to have considered that scenario?
Even consumer-grade products like ChatGPT can be interrupted while talking. We’re talking about an AI implementation, not some rigid set of
if...elsestatements.You talk as if they don’t already have customer facing implementations in effect, and that those all suck and don’t work worth shit. You also seem to think the “engineers” involved are custom doing anything and not just some sales person slapping a half baked product on every problem. When you get paid per deployed program, everything looks like a problem to solve with said program.
This is an issue of not having enough 911 dispatchers, due to the unwillingness to pay for them. The idea of putting in a chatbot, that can not even speed up diction let alone have any empathy is just wildly inappropriate in this situation. The fact is that these LLMs will (like in current deployments) most of the time have to pass the call to a person, drastically increasing time on the phone before action is taken. This is already seen in almost all LLM supported call centers, but a 911 dispatch is not a telecom company and more time on the phone is more death, injury and suffering.
That is the thought process of a child. If your defense of something with obvious issues is “Well I’m sure they know what they’re doing, because if they didn’t it would be bad”, I have some really, really bad news for you about the world.
If this was thought through by a capable adult, it would’t be implemented at all. The very fact that this solution is on the table is proof of the breakdown of rational thought. Or more likely, greased palms.
Your blind appeal to authority would hold more water if we haven’t already seen national adoption of AI products in emergency services that that have catastrophically failed to perform their basic functions, leading to real world harm. Audits of Flock cameras deployed in several cities found an error rate ranging from 33-70%. People are being arrested for things they didn’t do because the AI doesn’t work and nobody cares. People are being stalked because nobody thought of basic security controls.
Even worse, we’re talking about an AI implementation of if…else statements. Regular if else statements are consistent. As soon as you throw AI into the mix they become unpredictable. You introduce a failure mode that didn’t exist before that can and will have real consequences.
Yeah it’s definitely better to not answer all those calls
So… you do assume that you’re the first to think of this problem, or…?
If you assume that your brain is somehow superior[1] to everyone involved in the R&D, engineering, data analysts/QC, etc., then this may come as a surprise to you: You’re not even the first Lemmy commenter to think of it.
Literally multiple people here already thought about that and commented before you did. They’re not the first either; the author of the article herself even thought about it.
This idea is among the most obvious potential problems that come to mind within seconds or minutes of just hearing about the concept. It’s not a dumb idea by any means (it’s completely valid), it’s just not a particularly smart or special one.
911 is not the first department in that city[2] to use this tech, and that city isn’t even the first to adopt it. It’s silly to assume that such an easy-to-solve problem hasn’t been addressed by now, or that it somehow has not yet happened in other deployments.
To be clear: I make no assumptions that your brain is not superior to theirs either – I couldn’t possibly know either way. AFAIK, I’ve never even met you nor them… But I think it would be a safe bet that you don’t know either. ↩︎
BTW, have you heard cajun accents? There’s no way this bridge hasn’t already been crossed during their 311 trials. ↩︎
As someone who worked in the industry, your trust is wildly misplaced. I guarantee someone has already said internally this will not work and was shushed by a sales manager.
And as far as “I am sure they tested accents” goes? HA they likely noted it was a mess and moved on. You think the goal is a functioning system, but it is not it is just to make the sale.
Why does it have to be AI? Why can’t it be “press 1 for the highway car fire on 42, press 2 to be transferred to an operator”? Surely that’s much less likely to fail?
https://infosec.pub/comment/23116617
That doesn’t answer my question at all?
Ah, right, I suppose it’s contextual. You would also have to read the comment that it’s responding to. To summarize: it might not have to be AI.
However, 911 is a bit too dynamic for your solution to be viable as-is. For example, “Hey Siri/Google, I can’t move my hands. Call 911.”
Your idea is probably a decent starting point for brainstorming a solution — it just needs more thought, research, testing, etc. And I wouldn’t be surprised if a similar product already exists. (I’d actually be surprised if it doesn’t already exist.)
I believe the argument is that the probability of misclassifying an arbitrary segment of speech as the very particular answer “yes” is sufficiently low (with reasonably trained, current models) that we may ignore it under practical circumstances.
I find the non-zero probability of misclassifying in a little disquieting, but I suppose one way of looking at it is that this will probably increase the expected value of lives saved; the very probable case (calls are correctly filtered and a smaller proportion is passed through, allowing operators to respond to more new emergencies) may save a lot of lives, whereas the very improbable case (something that is not “yes” is misclassified as such) may endanger a few.
One thing that would worry me about just looking at expected values is the possibility of bias against a particular group of people or emergency type, but to me that seems unlikely in this case.
Edit: it also just occurred to me that if you are woefully unfortunate, you can probably just call again if you accept the assumption that there is a high probability the answer is yes given the model classified it as such. It might be more or less random chance?
Yeah I just don’t buy that there is a case where any “net gain” here is justified. These are people’s lives that you’re playing with by reducing them to what amounts to a balance sheet in the name of saving money.
The idea that you would create the possibility of denying someone emergency care that didn’t exist before, no matter how improbable (probably not even that improbable, given the propensity even the most advanced frontier AI models have for getting things wrong or hallucinating entirely despite simple instructions) just to save 10-15 seconds on average for other callers is not only absurd on its face, but morally bankrupt. You can personally ignore it because you don’t live there and it’s not the lives of you or your family at stake. I think those who this system could fail would not be able to ignore a flaw like that when it happens to them, and the idea that their peril was “highly improbable” will not be of much comfort to them.
The problem is that there are staffing shortages. The solution is hiring more staff. Trying to cheat our way out by implementing a system prone to unmitigatable flaws that could have life-altering or even life ending consequences isn’t a solution, it’s dystopian.
how long could the queue be? If someone can’t wait 60 seconds for the queue, but can take the 8 seconds to ask if it’s about X incident. I do imagine it will be smart enough to check their approximate location before giving the prompt