I'm not sure if I just thought this up sitting on the tiles in my shower room or if I dreamt it, but still.
Edit: I'm being cynical and I hate our blind faith in and reliance on LLMs.
At some point, it'll become acceptable to us when an AI shuts off life support for a patient at a hospital because it "calculated" the odds of survival
2 days ago by printf("%s", name); to c/showerthoughts
klankin 8 points a day ago
I think thats OPs point, Ai doesn't provide answers, it provides confidence.
So when they get it wrong 99% of the time, it doesn't actually matter cause we decided pleasing stakeholders is equivalent to actual success.
hoshikarakitaridia 22 points 2 days ago
This makes no sense.
Neural networks inherently work in a probabilistic latent space. The whole point is that we have a definable problem and we are unable to apply any deterministic mathematical calculations to it. That's why we call it a "black box problem" - we know what the solution looks like and we have a few examples of input and output data, but we don't know what the method looks like. In fact, even if we have an AI we don't know what the method looks like. If you wanna know more about this issue, look up explainable AI and you will find all the ways we have tried to come up with smart solutions like model heat maps to estimate their methodology. But these are all approximations.
Now if you ask an LLM what method it used to get to a solution, it will try to work backwards using its own reasoning and context clues to give you what its best estimation of the methodology is but that's not the same as its true methodology.
For me (and most people of the machine learning field I am aware of) an AI is inherently worse at calculations. That's what algorithms are for. And the only things sitting somewhere between an LLM and an algorithm are maybe Markov chains and randomforest. If there's anything remotely close to an AI model that is good at calculations, I am all ears.
webghost0101 9 points 2 days ago
I think your being a little to zoom in on the technicalities of modern day ai to tell the trees from the forest.
I am a bundle of connected neurons, effectively a neural network, i can calculate just fine.
Humans are “in general” horrible at doing math but thats besides the point of being capable of it.
Modern humans have a calculator in their pocket and we use it as a tool to make our own conclusions. Like ways “current day” ai often has acces to a calculator tool to do the same. Sure this is not ai doing the actual math but in practice its the same as a human with a calculator.
But the topic at hand is not even about current day ai but about hypotethical future ai that “somehow” is trusted enough to pull the plug on machines that keep a sick human alive.
Such future ai. If it would ever exist would have acces to such tools, possibly more complex tools to calculate odds of survival using (hopefully) a formula with a great number of input variables to consider all kinds of situations.
I would imagine such to be a known formula that humans could also independantly measure and calculate ending in a % chance of a survival. But somehow in this hypothetical future this would be a job given to ai.
So the question is less can ai solve math and more can we trust ai to interptet humam measuring tools to then put those values in a known formula tool to then act appropriatly on the results of said tool, and not at any point misalign/hallucinate.
hoshikarakitaridia 7 points 2 days ago
Ok but then why call it AI?
We don't have to dream about a distant future to find a thing that can do math for us. It's called a computer, and if it's in our home, it might even be a personal computer. In fact, we have them smaller and faster, and they have fancy designs and can communicate to the world but it turns out they got a small chip inside that does a bunch of arithmetic binary operations and then a lot of them also have a chip inside them that can quickly operate on huge matrices of values. We call those cpus and gpus respectively.
What I'm saying is, if it's just a threshold of different values, we were capable of making such a machine already, even before AI. And if it's an interpretative opinion, it turns out AI is notoriously bad at those.
So the only thing an AI actually helps with is to do something when we don't know how. But the issue was never how. Now there's something to be said about the predictability of death but there's an insane amount of compounding factors that contribute to this. An AI could maybe be trained on a subset of factors but the truth is, we don't have good input and output data, that's why AI won't help us here.
I'm not saying this to be pedantic, I'm making a point: even IF we decide that actually ethically it's fine for a machine to decide when someone has to die - which is quite a big leap - even then an AI won't be better than a human, so there's no point. And there's no way to make this look good, no way to sell this to the broader populus. I just don't see any way where this scenario is possible, let alone probable.
And I think that's important, because it's very easy to slide down this slippery slope, interpolate a time series forecast in our head and essentially conclude "it's going downhill so the future must be what our average dystopia predicted" and I don't think that's right or even helpful. There's quite a few things we should improve and our world is on the brink of a bigger societal change again, but change is never just bad. Change is an opportunity to improve the things that are important to us. And to do this, we need all the brilliant minds that see those issues to imagine those improvements, so we should make sure we know what needs to be done when we finally get the chance.
klankin 1 point a day ago
Thats really funny that your way for an AI to solve a problem is have something else solve the problem then claim ownership.
Like would it not be simpler in almost all cases to just use the tool directly?
Otherwise youre introducing an intentionally difficult to quantify number of errors.
klankin 1 point a day ago
Hell yeah I can be technically correct.
Under some definitions of intelligence, any transformation of information is intelligence, for example even 1+1=2.
So my math super ultra premium vbux AI discovery is the following python script:
print("2")
As per all test cases I've been able to conceive, this model has achieved a 100% success rate - and is definitively a major stepping stone to AGI (Is it now legally SGI in america?).
I'll expect my trillion in the mail.
klankin 10 points a day ago
My favourite dystopian theories are always the ones it turns out america has already been doing for 150 years.
Like the torment nexus ain't just inspirational, its systemically foundational.
Edit: Lmao someone else was creating credit score from first principals below
Diplomjodler3 12 points 2 days ago
It won't calculate the odds of survival. It'll calculate the revenue that can be extracted from you.
daannii 12 points a day ago
Have you ever seen this episode of Star Trek voyager?
The algorithm will be altered to say poor people don't deserve healthcare because they are more replaceable and their life has less value to the shareholders and billionaires. They will be denied medical treatment which will cause early death.
Radical_Socialist_t00t 7 points a day ago
So the same, just automated. If you believe capitalism cares about or values poor people right now well I've got bad news for you 😂
Crackhappy 7 points a day ago
redballooon 4 points a day ago
Why not? It makes some people nod "finally that's on screen", feeling a little not of relief, and then nothing changes.
Lushed_Lungfish 2 points 19 hours ago
That was an excellent episode of Voyager.
FrickAndMortar 11 points 2 days ago
It’ll never be acceptable, but they’re going to shove it down our throats anyway
farmgineer 10 points a day ago
it'll become acceptable to us
Speak for yourself. This is fully unacceptable.
Crackhappy 1 point a day ago
Your username + instance informs your comment.
HubertManne 5 points a day ago
not if will smith has anythihg to say about it.
ChicoSuave 4 points 2 days ago
Estimated Time to Death (ETD) feels like a near future stat used to justify killing terminal patients.
GroundedGator 4 points a day ago
And in a few decades we can start using Estimated Societal Contributions to determine how much a person's entire life might be worth and at what point it becomes financially irresponsible to continue providing any sort of support.
GroundedGator 1 point a day ago
I think that is a very optimistic number. The first 15 years are already a negative contribution.
klankin 2 points a day ago
I think we started that a few decades ago with credit score?
Like we don't directly kill anyone I suppose, but definitely indirectly increase the odds when they are less helpful monetarily.
Just for example with the gen Z suicide epidemic, by definition people are killing themselves not being killed. But to a company this is financially beneficial, cause a left-wing bias has on average resulted in more restrictive laws and a lower profit margin. It becomes an important business strategy to ensure they remain poor and anxious (and of course this causes some to be suicidal). Vice versa for boomers of course, elder care has never been better funded across human history.
Or another example, credit score companies would be "throwing money away" if they weren't also doing this for like union-oriented families, ie ensuring those who have statistically unionised (or follow any trends that might be too union-y), they're marked less valuable.
So they can kill people by letting them starve/OD/get in a car accident, or just shorten their lifespan with cheap housing, all without ever having to see a poor person.
Duamerthrax 3 points a day ago
Why not? Apparently, it's acceptable to use AI to decide bombing targets.
selokichtli 2 points 19 hours ago
TBH, I'd be surprised if it isn't happening somewhere already.
ArbitraryValue 1 point 2 days ago
Yes, because human doctors make a lot of mistakes.
paraphrand 1 point 2 days ago
Death panels, eh?
fluffy_cloud 5 points 2 days ago path: 0 26268418 26268887 26269105 26269533, hotness: undefined, score: 5, children: 0
klankin 4 points a day ago
I vote we should start flipping a coin for all decisions ever, to take the emotional pressure off everyone and blame the coin.
Cause I stg I'd kill god himself before we treat healthcare workers like humans, the only reasonable solution is the coin.
borinque808 2 points 2 days ago
What if it had? Do you think that makes it acceptable for the AI to decide to pull the plug (as per the OP's scenario)?
Rhynoplaz -3 points 2 days ago
Are you going to be the one to tell those other ten people's families that their loved one is going to die because another family wants to watch their brain dead Uncle breathe for a few more weeks?
borinque808 3 points 2 days ago
Actually, the conversation usually goes something like this:
Patient/family: do you have a kidney yet?
Doctor: no, not yet
(no mention of the family dealing with a difficult, personal decision about their loved one)
The larger question being presented is who gets to decide whether to pull the plug? Do you want such a decision about your spouse or child to be in the hands of another?
Rhynoplaz -2 points 2 days ago
Forgive my theatrics, just trying to put you in the mindset of 10 people that died because your human brain chose to state at a vegetable. No, probably not even that. They'll keep him on life support and go back to their own lives. And more and more people will die as he lays there alone and unaware.
Human emotion makes lots of bad decisions.
klankin 1 point a day ago
You think that can be accurately encoded in a text predictor?
Like sure eventually computers will be powerful enough to Turing machine an LLM into something useful - a la monkeys on typewriters. But the efficiency of such a system is compatible to natural selection in its probability distribution. When a Dr 40 years ago could have given you the same, and actually describe why without taking another 2^256 years to compute.
Rhynoplaz 0 points a day ago
I really don't understand why everyone thinks that chatbots and image generators are the only form of AI out there. AI has been around for decades longer than chatbots. Industrial AI isnt designed to chat. It's designed to work.
Rhynoplaz 0 points a day ago
I thought the spirit of the discussion was ethics of allowing machines to make medical decisions. Which, I think is a valid philosophical discussion that needs to happen soon, because the technology is quickly going to put us there.
Using a digital Magic 8 ball to make those decisions is just nonsense, and there's nothing to discuss.
username_1 -7 points 2 days ago
Blind faith? Reliance? Those exist only in anti-LLM fanatics' proclamations. Specialists who can use LLMs just... well... use it to save some time and effort where possible.
username_1 -2 points 2 days ago
You can start with licensing books usage. Imagine, now every idiot can open a book and read some crazy shit and nobody is going to stop that! Some people even learn reading outside specialized facilities! Outrageous!
AliOski 1 point a day ago
Not saying that the hypothetical situation of OP is impossible, some insurance companies will definitely try to do something like this and we, the people, should prevent anything like that to happen. However: even if such a system were implemented, it will be an AI trained specifically on the data related to health and life and death. It won't be a fucking LLM 😭 there might be an LLM on top of it to make the prediction more legible, but the actual prediction would be done by a specialized AI.
showerthoughts
@lemmy.world
44042
4160
5739
A "Showerthought" is a simple term used to describe the thoughts that pop into your head while you're doing everyday things like taking a shower, driving, or just daydreaming. The most popular seem to be lighthearted clever little truths, hidden in daily life.
Here are some examples to inspire your own showerthoughts:
- Both “200” and “160” are 2 minutes in microwave math
- When you’re a kid, you don’t realize you’re also watching your mom and dad grow up.
- More dreams have been destroyed by alarm clocks than anything else
Rules
- All posts must be showerthoughts
- The entire showerthought must be in the title
- No politics
- If your topic is in a grey area, please phrase it to emphasize the fascinating aspects, not the dramatic aspects. You can do this by avoiding overly politicized terms such as "capitalism" and "communism". If you must make comparisons, you can say something is different without saying something is better/worse.
- A good place for politics is c/politicaldiscussion
- Posts must be original/unique
- Adhere to Lemmy's Code of Conduct and the TOS
If you made it this far, showerthoughts is accepting new mods. This community is generally tame so its not a lot of work, but having a few more mods would help reports get addressed a little sooner.
Whats it like to be a mod? Reports just show up as messages in your Lemmy inbox, and if a different mod has already addressed the report, the message goes away and you never worry about it.
go to feed...
showerthoughts
@lemmy.world
44042
4160
5739
A "Showerthought" is a simple term used to describe the thoughts that pop into your head while you're doing everyday things like taking a shower, driving, or just daydreaming. The most popular seem to be lighthearted clever little truths, hidden in daily life.
Here are some examples to inspire your own showerthoughts:
- Both “200” and “160” are 2 minutes in microwave math
- When you’re a kid, you don’t realize you’re also watching your mom and dad grow up.
- More dreams have been destroyed by alarm clocks than anything else
Rules
- All posts must be showerthoughts
- The entire showerthought must be in the title
- No politics
- If your topic is in a grey area, please phrase it to emphasize the fascinating aspects, not the dramatic aspects. You can do this by avoiding overly politicized terms such as "capitalism" and "communism". If you must make comparisons, you can say something is different without saying something is better/worse.
- A good place for politics is c/politicaldiscussion
- Posts must be original/unique
- Adhere to Lemmy's Code of Conduct and the TOS
If you made it this far, showerthoughts is accepting new mods. This community is generally tame so its not a lot of work, but having a few more mods would help reports get addressed a little sooner.
Whats it like to be a mod? Reports just show up as messages in your Lemmy inbox, and if a different mod has already addressed the report, the message goes away and you never worry about it.
go to feed...
Calculating is exactly what AIs don’t do.
save