- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
You must log in or # to comment.
That title is written by someone who failed basic grammar. It should be “New Orleans will use AI instead of humans to answer 911 calls”.
What it says as it’s written is “AI will only answer 911 calls but not a human”… which doesn’t make any sense.Working my way through Person of Interest again. Just watched one where a search company was directing suicidal inquiries to ads instead of helplines. And that was before they get bought by the evil AI overlord. Its surreal seeing Palantir, Flock and all the tech companies race to implement a dystopia no sane person would ever want and there is so little push back. What the fuck is wrong with regular people?
New Orleans will kill its most vulnerable citizens to enrich tech billionaires.
there are tv shows that explain why this shit shouldn’t be done…
like AI rejecting calls because the people were (rightly) hysterical and the AI could not calm them down.
Sure, what could possibly go wrong?
This is exhausting y’all. God I fucking hate LLMs so much. This tech could be such a exciting break through, a step towards real artificial intelligence, but stuck in academia where it belongs. Instead of getting to be excited and interested this inanimate compilation of matrices is breathing down my neck constantly.
This tech could be such a exciting break through, a step towards real artificial intelligence
No, not really lol.
The best analogy for this that I’ve heard is that it’s like saying elevators are a step towards teleporters. Calling it “intelligence” is a marketing coup. It is not intelligence, artificial or otherwise.
But what is intelligence?
-every idiot who can’t tell the difference when they talk to a clanker
So now we’re at the stage of killing people directly, got it!
But universal healthcare leads to death panels or something…
You’re not wrong. Swatting in that area is going to be so easy. United States police force has already shown for years now that if AI says anything they’ll just go with it and not question it. So if someone lives in the area and you want them ‘dealt with’ by the police, it’s going to be very easy. And the police aren’t going to look into it or hold themselves accountable. And the AI won’t be questioned or removed. If this goes large scale you can just make an API… Don’t like someone on twitch, well when you know it there will be an app for that…
Bro what
All yall saying this is gunna kill people do not live in New Orleans. 911 is well know to move at the pace of a snail goin the wrong way up an escalator. This can only improve things (it won’t, but anything helps) , worst case it’s a change that no one notices. Some ones nephew got a contract for this, I garuntee it.
Oh I am sure putting a machine that’s only goal is to say things that seem real is going to help that…
Oh, how long until an AI ends up swatting people. There will probably even be different categories:
- Legit swat scenario, wrong address (or otherwise incorrect details make it turn tragic)
- Caller deliberately trying to swat innocent victim, AI goes along with it and adds credibility (not that humans seem to take away much credibility, so maybe this one shouldn’t really be a black mark for AIs if it does happen)
- AI just misinterprets a caller’s description, deploys swat team
- AI doesn’t trust the caller’s description, deploys swat team (anyone ever seen an agent start suspecting they were in an adversarial situation where something was injecting things into the context? I have)
- AI uses dispatching tool incorrectly, accidentally deploys swat team
Or just as likely never dispatches in the case of a real emergency. I have a hard time thinking of what could go right.
Hire more operators? Nah Ai bot? Sure Solid logic I also love how bloated police budgets are but get the other departments in the city related are struggling to keep the lights on.
Everyone should read before reacting:
This is intended to handle the increase of calls that are related to one incident, so callers get automatically routed to an AI agent who asks if they are calling regarding the incident. If the answer is yes, then callers can receive information or updates, and if it’s no, then the callers are transferred to a human.
This makes sense. A big thing happens that you are aware of. Geolocated calls from that area flood in. An operator, who could be getting useful info, has to handle a flood of “Guess what I just saw!”
Note that I think they should spend the money on more operators instead, but at least this seems like it has good guardrails.
And if someone tries to call with new unreported information about an already reported incident? Decision tree or geolocation routes caller to recorded information. No way of knowing if information is important or not without talking to the caller first. Possible downside is far greater than possible benefits.
Edit Multiple calls about shots in neighborhood, if victim is last to call are they forced to listen to an update before triage allows transfer? No sir I don’t like it.
Maybe this is an improvement in this location, but as many have said a staffing issue.
Functional societies have costs, often those costs are far less than dysfunctional societies.
Except that could be done without AI. Just have a recording that says "we are aware of … taking place near you, are you calling about this ? “, then answer with yes/no.
No need for bullshit AI.
Have you ever met humans? That would be a recipe for “no I’m not calling about that, I want a human” just to get to a human and say you’re calling about exactly that.
My understanding from other articles is caller satisfaction is greater with the TTS system, and it uses voice IVR to triage of they have additional information. AI is not bullshit in and of itself. This product seems barely diffeent from what they provided with APEX.
Recordings also can’t auto respond with updates to the caller like this can, which is a pretty nice feature, even if it’s just TTS.
Can you share articles, interested how design a dynamic IVR to handle unfolding event and update prompts in realtime as new information is added to original incident.
caller satisfaction is greater with the TTS system,
You know what’s great about the valley callers who don’t make it?
No bad survey results!
My aunt was pissed when she found a hospital satisfaction survey addressed to her mother, a few weeks after her mom died at that hospital.
Even assuming perfect conditions - where the caller very clearly states what’s happening and the AI agent understands everything properly - this is a waste of resources. Just hire more operators dawg. It’s really not that complicated.
And then think about what happens outside of those perfect conditions. So pretty much most calls.
You don’t need AI for that. Automated phone answering systems do exist, and can redirect people.
This is basically a beefed up IVR with better customer satisfaction reports, so… yes.
But as I said in the other comment, giving updates to callers is something a standard answering system can’t do.
How many operators to handle say 100 calls in 5 minutes? Now imagine the cost to staff a small 911 center to ensure your can handle those calls 24/7. The money doesn’t exist.
Butts in seats don’t exist right now either. As long as the fancy ivr allows the caller to reach a human with minimal effort, this seems like a scary sounding but neat
assistantsolution to a problem without a real solution.Well it wouldn’t cost much cause you’d only need to employ them for a very short amount of time on occasion. Just have a large team of trained staff on call - its just that simple!
Listen to yourself. Have a large team of trained staff on call. How much do you think that costs? No go on guess. Do some calculations. The people do not exist to fill the seats. The money does not exist to pay the non-existent people. This is an attempt to make an already shitty situation better.
I work in this industry and I read the article. I understand what they are doing. I hate AI but this seems reasonable.
yeah no shit…
So did I completely misread your comment? LOL
So, say a tower block fire happens at 2 am. You’re getting 50 calls a minute, 20% from residents and the rest from people calling to say they see a fire.
How long would it take for your trained staff to spring into action compared to an on demand AI? And are you paying your on-call staff to wait, or only when they’re on the phone?
edit: I’m thinking about the Grenfell Towers. I listened to the BBC’s coverage of the enquiry, and it was rough. Gave me a new appreciation for what emergency responders do.
Removed by mod
Thank you for your insight. I had logistical questions I thought you might could answer, but I see now that I was wrong about that.
I wish this were pinned at the top. It seems like almost nobody reads beyond the headline.
This is one of the few actually cool and useful implementariins to me. It reminds me of my work writing systems to detect antimicrobial issues in hospitals.
A shame it’s getting so much hate, but they did choose to brand with “AI” so they get what they ask for.
Joe’s Apartment, the 911 call center be like
Hey I bet Kanye West and a bewildered Mike Myers can explain why…
…too bad Kanye West isn’t around anymore
You are absolutely right to call me out on that. He did break in, and he is now stabbing you. I will adjust for this in the future. Is there anything else I can help you with?
Brilliant. I can’t wait for all the false calls they will generate with this tech too and then use it as excuse to Gestapo and sent people to the prison camps.
Can I get desperately needed help please??? N.O.
People are going to die because of that, even if AI is merely handling call “triage”.
I love it.
All of the movies and novels warning of danger from AI, news of AI convincing people kill themselves, suggesting ‘remedies’ that would cause someone to die, or going wrong in some other way… and they decide it would be a great plan to have it answer emergency calls from people who could be in the process of being stabbed to death.
Why? Because it’s cheaper than hiring enough people to do that job.
The nature of the job means that’s very literally deciding that it’s better for more people to die than to increase the budget for 911 operators.
Even more than this one instance, it’s the same reason that AI is running pretty much everything and the streets are full of AI-powered robots when things go wrong in all of those movies and books. Because it’s cheaper than paying a person to do it.
I can’t think of any examples even in science fiction stories where the richest people in the world create AI and robotics and manage to automate pretty much all human labor in order to cut costs that end right there with “and then the super rich now in direct control of all production and society divided their wealth and power evenly among all humans and everyone lived happily ever after.”
That’s just to unrealistic, even for fiction.
This will kill people.
Yes. So many. Everyone involved in this decision, on both the legislative and corporate sides, needs to get the fucking wall.
But it’s new Orleans, so some of the difficult to emotionally comprehend number of deaths it causes might at least be funny.
I have mixed feelings on this and how they are implementing it. Here’s the direct quote
This is intended to handle the increase of calls that are related to one incident, so callers get automatically routed to an AI agent who asks if they are calling regarding the incident. If the answer is yes, then callers can receive information or updates, and if it’s no, then the callers are transferred to a human.
In a way this will reduce caller volume for the same incidents being called in over and over again. Seems like they are using a direct yes or no question which decreases any issues with the AI making shit up.
I do see how this can free up humans to deal with other calls and dispatch lifesaving services faster, eliminating repeat calls.
However, even if the AI works as described (which is a big if anyway), having to navigate even a simple yes or no will reduce time to lifesaving care.
Our 911 system has been slow and inadequate for a long time now. There has been an absolute refusal to fix it, and this has led to deaths. I do see this application easing that issue a bit, but I also see how this will cause deaths in a different way. Is it better to die on hold waiting for an operator or is it better to die navigating an AI system? I think less would die overall with the new system. It has been trained using the non-emergency call line.
To be clear I want us to just fund people to staff the 911 call center, but these shortages have been an issue for decades. Local governments refuse or are unable to take on the additional costs. I’m really unsure on how to feel about all this. My gut reaction is disgust, but I do see what they are trying to do. I guess we’ll have to wait for the data and stories
The key is to remember you can’t actually fix anything except with the wacky fashionable new scam. .and since when does a y/n require fucking “ai"?
I think the summary of the existing incident and providing updates is where they are using the AI. Which if given a limited data set should work.
These LLM models do actually a good job of summarizing a list of facts you feed it. The issues happen when you want it to make something “new”
For 911 I’d rather the government find the money to fund them, but I do feel like they are using AI in an appropriate scope.
AI should not be used in life and death situations like a 911. But the way they are going about the roll out is demonstrating how to actually use AI as a tool and not some miracle machine.
They should have left it only on the non-emergency line and called it there.
A person could do this better faster cheaper lower odds of failure.
‘It can kinda do it sometimes so it’s ready to go live’ is slopper propaganda. There is no “AI” extant at this date.
This is just an excuse to kill poor people. As all “ai” is.
Yeah, I can just imagine how well a “AI” would deal with a situation like this when they can’t even operate a drive-thru properly.
I ask this out of genuine, burning curiosity, and mean no rudeness by this…
Did you read the article that the OP linked to?
I don’t mean to single you out with this question, I’m just confused after scrolling this far through sooo many similar comments. It almost feels as if I somehow ended up reading a different article from everyone else.
Did you read the article?
callers get automatically routed to an AI agent who asks if they are calling regarding the incident. If the answer is yes, then callers can receive information or updates, and if it’s no, then the callers are transferred to a human.
So if somebody calls and is completely out of his mind, what will AI do?
So if somebody calls and is completely out of his mind, what will AI do?
Is this a rhetorical question?
Just in case you wanted a real answer: The caller is obviously not saying “yes,” so transfer to human.
To be clear: I don’t know anyone at Carbyne or OPCD. I can only offer speculation. But I’m curious: What would you have guessed?
I guess you don’t have lots of interaction with AI yet?
What if someone is obviously out of their mind and says yes? Random example, say someone has a stroke and has just barely figured out how to use the big red emergency call button on their phone screen. The AI asks if they are calling about blah blah emergency. Maybe the person only understands half of it, thinks they’re being asked if they’re calling about an emergency and say “yes”. Now they’re stuck having a form letter read to them by AI while their brain is dying. A person would catch something was up even if they said yes.
The article even notes this can be a problem without anything medical going on:
It’s also possible that AI agents handling 911 triage could fail to understand individuals with stronger accents and dialects, as well as differences in pitch and articulation. This is because AI is programmed with automatic speech recognition systems and, therefore, could potentialy be unreliable when presented with a speech unlike the ones used the AI is trained to recognize.
Responding to people having emergencies is too important a task to leave up to voice recognition slapped on top of an LLM.
I don’t know the concrete answer to your "what if,’ and I don’t think you or the author of the article do either. That would be a question for Carbyne.
I would speculate that there is an answer, and that neither you nor the author are the first to think of these problem scenarios. There’s design, engineering, and real-world testing involved.
“Well I just trust they thought of it” isn’t good enough.
But…just think how many money they can save! You are being selfish!
/s
They know it will. They don’t care.
deleted by creator
And then it will be nobody’s fault, nobody could have predicted this. <- Their pre-planned excuse.
