A.I. danger

feedmpenzaandhewillscore said:
Ai can spot cancer growing on scans better than any trained oncologist can. What it could do across our buckling NHS and care services where there simply isn’t enough people to make it work could be revolutionary in the coming years with some of the robots that are being developed by Musk etc.

Flip side that needs looking at is how it’s already collapsing the job market. A lot of entry level roles are vanishing and it’s going to eat into potential apprenticeship roles when we are seeing rising unemployment of grads etc. what happens when driverless cars, vans, hgv and fully automated warehouse and distribution centres go online? A lot of the world’s population will be deemed useless for work? What happens to these people, I think it’s a far more pressing matter at the moment than if eventually AI will wipe us out.
Click to expand...

Funnily enough, I said around 18 months ago on BM that the only jobs that will be meted out will ones of Human to Human contact and some scoffed at the notion. Luckily for me, I'm in one of those jobs.

I, even, said to my daughter yesterday as she's planning to move to the drug counselling field, that this will escalate by the time she's done all her qualifications and set up and there'll be plenty of work.

Idle people seek a release from themselves. I don't mean that as a slur, just that when one is doing nothing, it's easy to slip into dependency to escape 'this life'. People will feel 'what is the point?' when the rich are going to be dividing more of the haves and have nots and there's nothing for Humans to do and the trend in today's young is not to do too much as it is.

Interesting times, really.
 
There's one thing I haven't even contemplated, but the future will have kids taught by AI! There'll be no need for teachers at all and 'history' (as much as it's now being taken apart by likes of Trump wanting to reframe American history) will be tailored more and more.

Sam Altman also had a discussion on his new GPT 5 model with Cleo Abram...

 
Bigga said:
There's one thing I haven't even contemplated, but the future will have kids taught by AI! There'll be no need for teachers at all and 'history' (as much as it's now being taken apart by likes of Trump wanting to reframe American history) will be tailored more and more.

Sam Altman also had a discussion on his new GPT 5 model with Cleo Abram...

Click to expand...


On one side you have the liberals coughing up all sorts of shit and on the other you have Americas Putin/XI, makes you wonder where all the normal people went?

Mind you when you put a walking corpse and a cackling idiot in the firing line as the best of the rest you get what you deserve. Maybe AI can be an improvement who knows :)
 
RadcliffeRick said:
The worrying thing is that we just do not know if anything is genuine any more:-(
Click to expand...

We really don't. And if people are saying "AI isn't very good so it's easy to spot", then you're not looking hard enough. AI bot manipulation is everywhere online these days. And I mean, everywhere. I don't trust any comment sections posted on political topics on major platforms - there are people out there trying to manipulate the public's opinions in a really sinister way. Trying to make you think that "most people think like this" or insert counterarguments which sound reasonable and plausible on the surface but are actually misinformation.

I was on Reddit yesterday, just browsing, and I ended up on a Subreddit which I'll describe as politically left (some might call it populist or far left) which has recently grown quite quickly. I'm a yoghurt-knitting centrist, so I like to keep up with what's going on across the spectrum.

One of the posts in there was about Land Value Tax. It was full of seemingly lively debate about the pros and cons, but I noticed a lot more than half of the comments were quite critical of the idea - which I just found weird because it's a left wing Subreddit, and I know LVT is usually quite popular with the left.

Then I looked a little closer, here is a selection of the usernames from that thread - see if you can spot a pattern:

EnoughPsychology6432
Striking-Lion9024
Positive-Relief6142
Normal-Ear-5757
Fun-Aardvark-7783
Vegetable_Grass3141
Strong_Remove_2976
FewEstablishment2696
Alive-Turnip-3145
Particular-Way-8669
Odd_Government3204
RemarkableFormal4635
SecondSun1520
Altruistic-Bat-9070

Here is the thing, though - they weren't all flat out disagreeing with it - they were planting seeds of "maybe this is a bad idea". These bots were starting conversations where some of them would start off with a positive opinion that was badly thought through, and another bot would debate them in a thread where they seemed to counter all of their points effectively. If you follow these threads, 90% of them seem to end in "LVT is a bad idea". Then I noticed a few of the moderators for the Subreddit also had names with the same schema as the above.

I'm not even necessarily for or against the idea myself, but I find stuff like the above extremely worrying. People are quite clearly using AI to create entire sock puppet communities that feel "alive" which are designed to skew public perception on individual topics and ideas. And this isn't necessarily a left v right thing, I have seen examples of this on both sides, but it is definitely skewing towards whatever opinion would benefit either the wealthy or those that seek to divide us.
 
It will kill us all, but it will go from completely harmless to extinction level event danger so quickly that we won't even know its happening until it's already happened.

So it's not all downsides I guess.
 
feedmpenzaandhewillscore said:
Ai can spot cancer growing on scans better than any trained oncologist can. What it could do across our buckling NHS and care services where there simply isn’t enough people to make it work could be revolutionary in the coming years
Click to expand...
The problem comes not when we start using AI to help the oncologist, the problem comes when we decide we don’t need to oncologist any more. Big tech are already doing this with content moderation, and it can’t tell the difference between someone posting a picture of their kid doing gymnastics and child abuse images.

As others have said, the danger with AI isn’t that it will be so good that it replaces everyone, it’s that it’ll be “good enough” to replace a lot of jobs leading to much shittier services overall.
 
Bigga said:
There's one thing I haven't even contemplated, but the future will have kids taught by AI! There'll be no need for teachers at all and 'history' (as much as it's now being taken apart by likes of Trump wanting to reframe American history) will be tailored more and more.

Sam Altman also had a discussion on his new GPT 5 model with Cleo Abram...

Click to expand...


I’d seen this one mentioned somewhere before. Imagine a tutor specifically designed per student to work with them in their own personal learning style and help around any leaning difficulties like dyslexia etc.

I don’t think they will phase out teachers but will be teaching assistants.
 
grunge said:
I’d seen this one mentioned somewhere before. Imagine a tutor specifically designed per student to work with them in their own personal learning style and help around any leaning difficulties like dyslexia etc.

I don’t think they will phase out teachers but will be teaching assistants.
Click to expand...

I hear you, but why bother with a teacher at all? You can programme AI to deal with dyslexia and the pupil won't feel the human emotion of embarrassment/ inadequacy to software/ a robot.

I was going to pop this comment in separately but here is just as good. I know many think AI will "kill us all" but it won't, it will just make Humans more docile and less maverick in nature.

A future generation will just have robot companions/ aides with them at all times, I think. I think AI will assimilate into Human life and they'll tend to monitor Humans for mental, emotional and physical changes. They may be small pocket sized companions that 'put themselves away' at times to recharge during their own Human's sleep time or even be mini hovering almost drone-like machines.

Humans will either shrink in population with Gov control or explode because there's not much else to do with most jobs taken by AI.

Anything could happen.
 
I think it is pretty clear people don't actually understand what Artificial Intelligence is.

As opposed to, advanced programming and general automation. Which is where we are just now really, or at least as is available to the public.
 
Damocles said:
It will kill us all, but it will go from completely harmless to extinction level event danger so quickly that we won't even know its happening until it's already happened.
Click to expand...
not that i particularly disagree with your forecast,
but how can it be possible to know it's already happened if we have become extinct from it?
 
Damocles said:
It will kill us all, but it will go from completely harmless to extinction level event danger so quickly that we won't even know its happening until it's already happened.

So it's not all downsides I guess.
Click to expand...

You’ve been saying that for years now and will probably be saying it on your deathbed.
Might happen but I say very unlikely. Who knows.
 
Bigga said:
I hear you, but why bother with a teacher at all? You can programme AI to deal with dyslexia and the pupil won't feel the human emotion of embarrassment/ inadequacy to software/ a robot.

I was going to pop this comment in separately but here is just as good. I know many think AI will "kill us all" but it won't, it will just make Humans more docile and less maverick in nature.

A future generation will just have robot companions/ aides with them at all times, I think. I think AI will assimilate into Human life and they'll tend to monitor Humans for mental, emotional and physical changes. They may be small pocket sized companions that 'put themselves away' at times to recharge during their own Human's sleep time or even be mini hovering almost drone-like machines.

Humans will either shrink in population with Gov control or explode because there's not much else to do with most jobs taken by AI.

Anything could happen.
Click to expand...

Im thinking in terms of Now, or in the next couple of years. I feel that teaching needs a personal touch that cannot yet be replaced with an LLM. a teaching assistant being a fine tuned LLM could well be done soon. I think power/compute is a limiting factor on that one a the moment. as well as knowing how we'd need to fine tune those models to work the best per child.

For teaching an LLM would have to be a pretty large one to help limit hallucinations and fine tuning those isn't cheap

A guy where i work fine tuned Deepseek to put back all the data that the Chinese Gov censored out and that took quite a bit of prep work with our local hardware and a few hours and at least a few hundred dollars worth of cloud compute to do, so fine tuning per child is currently quite prohibitive on that front. Maybe something like this would start out with maybe 10/15 variants and then expand from there.
 
I wonder what the implications of driverless cars are. Presumably the first monumental adoption of AI over human? Personally, im looking forward to not driving, its become hideous, too much traffic, too many rules and cameras and ridiculous signage and arseholes and crazy car parks.. i could go on. BUT... who pays the fines, and what happens when the system tells us we don't need to do that journey or we haven't built up enough 'credits' to take that journey or that journey is better taken later. What happens if the system conspires with institutions to bring us in to pay a bill or arrest us if we are agitating to change the system or doing too much workers rights commentary online. Again.. all very "Minority Report" but these genius authors like Philip K Dick were geniuses for a reason. The trouble is, by the time adoption or roll out is 80% there..its too late!
 
grunge said:
Im thinking in terms of Now, or in the next couple of years. I feel that teaching needs a personal touch that cannot yet be replaced with an LLM. a teaching assistant being a fine tuned LLM could well be done soon. I think power/compute is a limiting factor on that one a the moment. as well as knowing how we'd need to fine tune those models to work the best per child.

For teaching an LLM would have to be a pretty large one to help limit hallucinations and fine tuning those isn't cheap

A guy where i work fine tuned Deepseek to put back all the data that the Chinese Gov censored out and that took quite a bit of prep work with our local hardware and a few hours and at least a few hundred dollars worth of cloud compute to do, so fine tuning per child is currently quite prohibitive on that front. Maybe something like this would start out with maybe 10/15 variants and then expand from there.
Click to expand...

Far enough on your teaching position now it's better explained along the timeline.

And I, much, prefer Deepseek to other models.

Not sure why.
 
SkyBlueFlux said:
We really don't. And if people are saying "AI isn't very good so it's easy to spot", then you're not looking hard enough. AI bot manipulation is everywhere online these days. And I mean, everywhere. I don't trust any comment sections posted on political topics on major platforms - there are people out there trying to manipulate the public's opinions in a really sinister way. Trying to make you think that "most people think like this" or insert counterarguments which sound reasonable and plausible on the surface but are actually misinformation.

I was on Reddit yesterday, just browsing, and I ended up on a Subreddit which I'll describe as politically left (some might call it populist or far left) which has recently grown quite quickly. I'm a yoghurt-knitting centrist, so I like to keep up with what's going on across the spectrum.

One of the posts in there was about Land Value Tax. It was full of seemingly lively debate about the pros and cons, but I noticed a lot more than half of the comments were quite critical of the idea - which I just found weird because it's a left wing Subreddit, and I know LVT is usually quite popular with the left.

Then I looked a little closer, here is a selection of the usernames from that thread - see if you can spot a pattern:

EnoughPsychology6432
Striking-Lion9024
Positive-Relief6142
Normal-Ear-5757
Fun-Aardvark-7783
Vegetable_Grass3141
Strong_Remove_2976
FewEstablishment2696
Alive-Turnip-3145
Particular-Way-8669
Odd_Government3204
RemarkableFormal4635
SecondSun1520
Altruistic-Bat-9070

Here is the thing, though - they weren't all flat out disagreeing with it - they were planting seeds of "maybe this is a bad idea". These bots were starting conversations where some of them would start off with a positive opinion that was badly thought through, and another bot would debate them in a thread where they seemed to counter all of their points effectively. If you follow these threads, 90% of them seem to end in "LVT is a bad idea". Then I noticed a few of the moderators for the Subreddit also had names with the same schema as the above.

I'm not even necessarily for or against the idea myself, but I find stuff like the above extremely worrying. People are quite clearly using AI to create entire sock puppet communities that feel "alive" which are designed to skew public perception on individual topics and ideas. And this isn't necessarily a left v right thing, I have seen examples of this on both sides, but it is definitely skewing towards whatever opinion would benefit either the wealthy or those that seek to divide us.
Click to expand...
philfodenswristsock said:
I wonder what the implications of driverless cars are. Presumably the first monumental adoption of AI over human? Personally, im looking forward to not driving, its become hideous, too much traffic, too many rules and cameras and ridiculous signage and arseholes and crazy car parks.. i could go on. BUT... who pays the fines, and what happens when the system tells us we don't need to do that journey or we haven't built up enough 'credits' to take that journey or that journey is better taken later. What happens if the system conspires with institutions to bring us in to pay a bill or arrest us if we are agitating to change the system or doing too much workers rights commentary online. Again.. all very "Minority Report" but these genius authors like Philip K Dick were geniuses for a reason. The trouble is, by the time adoption or roll out is 80% there..its too late!
Click to expand...
A.I. Will be brought into the picture more and more specifically by the “Authorities” to better control what the Plebs are doing -they simply can’t help themselves -so on that basis I’m against it
 
You must log in or register to reply here.
Unread Posts

Don't have an account? Register now and see fewer ads!

SIGN UP
Back
Top