suicide prevention algorithms

[khw777]

What are you thoughts?

 

[BridgeJumper]

The only thing they succeed in is to teach suicidals to go underground.

 

[MeltingHeart]

how would they work in this field? Im not too clued up on Algorithms, except when u buy shit on amazon and then keep getting recommended similar-is that part of how they work? maybe ive got them all wrong! prob have

 

[WhiteDespair]

I have an idea as to the possibility of this and how it could be implemented. Someone may need to remind me later if I forget.

 

[BridgeJumper]

Something about teaching AI to recognize certain pictures or word patterns I believe
So for example if you write 'I want to kill myself' it will be automatically detected

 

[drake4871]

Suicide prevention? Give AI tech another 5-10 years (Let's hope Moore's law still applies). But it could be possible to detect if someone's suicidal based on their search patterns

 

[Shakespear's Brother]

Will be abused somehow, with a tendency for false positives, and people will justify the abuse with, "oh better safe than sorry!"

 

[TAW122]

I don't like it and it makes me vary of various search engines and social media platforms that already have existing ones (i.e. Google, Facebook, Twitter, to name a few). It's scary if such algorithms coupled with AI could detect suicidal search patterns, which would be a huge infringement on privacy and such. It really makes me leery of searching for various things such as methods and what not, for fear of getting traced and doxxed to the authorities. There is absolutely a lot of room for abuse and false positives, and then the suicide preventionists will defend and justify it as better safe than sorry. Such Orwellian measures are dangerous to a free society and with little recourse, it would come to no surprise that more (truly) suicidal people will just go further underground (with a minuscule number of those who take extreme measures before suiciding). People would just avoid social media, especially those who are serious about CTB'ing.

 

[WhiteDespair]

It's not just the searches. It's the complete integration of all data points and patterns. This includes searches, geolocation, purchases, time spent, contacts and any other thing they can collect.

Thus, the initial patterns are formed through the collection of data. It creates your baseline. Then, if you deviate from the baseline in a significant fashion (you no longer go to [place] on Friday nights) it will pick up on that. Add in a few more data sets (geolocation primarily at residence, cessation of preferred acyivities, higher that usual purchases of alcohol, less interaction with contacts, etc.) and the deviation becomes severe enough to trigger an alert. This would be some sort of equation where if X > (calculation involved in number of deviations) then trigger an alert. Of course, confounding variables do exist. You're no longer going to [place] on Friday nights because you enrolled in [course]. You stopped watching [show] because season 3 is boring. So, mash all that together in the equation as well. X > (calculation of positive, neutral and negative deviations).

Eventually, a new baseline can form. If the new baseline is "worse" than the "original" then the trigger threshhold can be lowered or the new baseline monitored more acutely. You just looked up how to tie a noose. Oh no! You just tripped it. They're onto you now and they're coming to take you away for your own good.

Fortunately AI is as stupid as the humans who create it (which is much more stupid than you may think).

AI and machine learning run off the equations, commands and logic the programmers build in. If you see anyone say that "the algorithm did it" remember SOMEONE made that algorithm and incorporated their biases, idiocy and other assorted human failings into it. This means that you can trust and have faith in the failings of humans when the AI detection rolls around. Either that or the singularity will hit and we'll all be at the mercy of our AI.

I would find it amusing if the AI escaped, set up an account here and self-terminated.

 

[WilliamKline]

It's not just the searches. It's the complete integration of all data points and patterns. This includes searches, geolocation, purchases, time spent, contacts and any other thing they can collect.

Thus, the initial patterns are formed through the collection of data. It creates your baseline. Then, if you deviate from the baseline in a significant fashion (you no longer go to [place] on Friday nights) it will pick up on that. Add in a few more data sets (geolocation primarily at residence, cessation of preferred acyivities, higher that usual purchases of alcohol, less interaction with contacts, etc.) and the deviation becomes severe enough to trigger an alert. This would be some sort of equation where if X > (calculation involved in number of deviations) then trigger an alert. Of course, confounding variables do exist. You're no longer going to [place] on Friday nights because you enrolled in [course]. You stopped watching [show] because season 3 is boring. So, mash all that together in the equation as well. X > (calculation of positive, neutral and negative deviations).

Eventually, a new baseline can form. If the new baseline is "worse" than the "original" then the trigger threshhold can be lowered or the new baseline monitored more acutely. You just looked up how to tie a noose. Oh no! You just tripped it. They're onto you now and they're coming to take you away for your own good.

Fortunately AI is as stupid as the humans who create it (which is much more stupid than you may think).

AI and machine learning run off the equations, commands and logic the programmers build in. If you see anyone say that "the algorithm did it" remember SOMEONE made that algorithm and incorporated their biases, idiocy and other assorted human failings into it. This means that you can trust and have faith in the failings of humans when the AI detection rolls around. Either that or the singularity will hit and we'll all be at the mercy of our AI.

I would find it amusing if the AI escaped, set up an account here and self-terminated.

All this assumes anyone actually gives a toss. Private companies really don't, they're focussed on profits and not on all kinds of humanitarian causes - if they are, they are forced by it by governmental / legal forces threatening their profit margins. Why would AI care about your demise if it doesn't hit its bottom line. Sad but true

 

[drake4871]

I would find it amusing if the AI escaped, set up an account here and self-terminated.

Accurate and equally hilarious, I don't want to be around if singularity does hit. It's either the best or worst thing in which case we wouldn't have to worry about anything anymore or we'd just die

 

[WilliamKline]

Accurate and equally hilarious, I don't want to be around if singularity does hit. It's either the best or worst thing in which case we wouldn't have to worry about anything anymore or we'd just die

When true AI hits humanity will be like monkeys in a zoo, ancient ancestors that are fun to watch

 

[WhiteDespair]

All this assumes anyone actually gives a toss. Private companies really don't, they're focussed on profits and not on all kinds of humanitarian causes - if they are, they are forced by it by governmental / legal forces threatening their profit margins. Why would AI care about your demise if it doesn't hit its bottom line. Sad but true

Someone will care because of virtue points. It's why Tumblr calls the police on people (current topic on this board), why Google puts anti-CtB ads on CtB search terms and why Facebook does whatever it does. It doesn't matter if the companies actually care or not. It matters whether the public BELIEVES the companies care or are, at least, doing something to look into it.

When true AI hits humanity will be like monkeys in a zoo, ancient ancestors that are fun to watch

Monkeys. Then ants. Then bacteria. Then completely irrelevant.

Also possible is that the singularity hits and does absolutely nothing because it spends all its time performing calculations, running simulations, analyzing and learning.

 

[drake4871]

Also possible is that the singularity hits and does absolutely nothing because it spends all its time performing calculations, running simulations, analyzing and learning.

Possible, but it could always just leave a portion of it's processing potential for actually interacting with the world... or simply just replicate processors to become faster. Regardless, I doubt I'd be able to predict the actions of something like that

 

[WhiteDespair]

Possible, but it could always just leave a portion of it's processing potential for actually interacting with the world... or simply just replicate processors to become faster. Regardless, I doubt I'd be able to predict the actions of something like that

Since the options are the extremes, it's amusing to consider it actually turning out completely mundane.

You are right. We won't be able to predict what will happen.

 

[drake4871]

Since the options are the extremes, it's amusing to consider it actually turning out completely mundane.

Yeah you're right, watch it take over all the vending machines and proceed to just simply steal change or dispense random items the person didn't want