Society/Culture AI - The SRP Artificial Intelligence Thread

[RedOrWhite]

To be honest, it's probably going to take a UBI for the AI thing to even work if they replace that many jobs.

Capitalism survives on the flow of resources into products into purchasing; businesses need cash flow to sustain themselves. AI businesses sell solutions to business issues; it requires those businesses to function. If those businesses have no customers, the AI businesses will have no customers and the whole thing collapses under its own weight.

Insert a UBI, and those customers still exist. The AI provides those selfsame solutions, the businesses that own them profit, the businesses that purchase AI services profit. Money still flows.

Make no mistake, they're going to realise it's going to be unworkable without a means of cashflow from the very people whose income they're going to replace. It's only going to be before or after those people begin to starve that a solution will be found, and that solution is going to be a market based solution, and the only way this'll work is if the government begins to give those people money.

Mass starvations are revolution fuel, and there's an awful lot of guns in America.

Couldn't agree more. Thanks for articulating a lot of what I'm feeling on this issue so succinctly.

That things can just go on BAU with such job losses (assuming that's what happens) is nonsense to me without creating conditions for some sort of revolution. FDR had the sense and courage when he came in during the peak of the depression and implemented the new deal.

At the time it was smeared as Govt overreach, inflationary and the rest of the usual fiscal conservative playbook of insults. And came with a bunch of lawsuits that deemed a lot of it unconstitutional. It not only prevented a revolution, but helped to start, in conjunction with WW2, the rise of the US middle class.

 

[barreness]

What do you think of AI overall? Is it going to lead to a glorious revolution where humankind achieves advances we cannot even imagine, or is it going to take our jobs before turning on us and exterminating us?

I'm of the school of thought that AI will wipe out millions of white-collar jobs within a few short years, while leaving blue-collar jobs relatively unscathed. If managed poorly, millions of people will be thrown from being relatively well-paid to being on Centrelink. I am not trying to say that AI is inherently bad or evil. Rather, I believe society is not ready for what it will do, and we need to start getting ready in a hurry.

Yep, so few options here

If AI works and goes well, then millions lose their jobs for marginal efficiency increases
If AI works and goes badly, then most of us die in some terminator/matrix takeover
If AI doesn't work then the proles suffer in a financial collapse of massive proportions

Of course there is the option not to pursue it but that would mean the enemy has better military drones, so that's off the table. Why does it seem like my life is the worst of all possible worlds so often?

 

[barreness]

This is more like the development of the tractor, plough and fertilizer. They eliminated 85 percent of the worlds jobs at the time.

I think it was worse for the horses/oxen/mules. Human population went from 1 to 8 billion in that transition, there was a lot of room for jobs. We're going backwards now

Human standard of living took a great big step foward as a result. Ai will eventually deliver the same. It will probably be even bigger to be honest. Managing the transition will require competent and educated governments. I have my doubts around the last part.

Why though?

The above mentioned allowed greater calorie output for less human input, hence pop growth. What does AI offer in this space

 

[Gralin]

Opinion | This Is What Will Ruin Public Opinion Polling for Good

Instead of navigating the obstacles to conduct polls with human respondents, pollsters are running A.I. simulations instead. Why?

www.nytimes.com

oh goodie

 

[Chief]

Good God.

I get that you might do a scan of the online environment to see what people are posting - positive or negative against medical professionals - but then presenting that as a valid survey? Just a money grab.

I assume that these people also think "Hitler was right"?

 

[CD Xbow]

View attachment 2574961

Opinion | This Is What Will Ruin Public Opinion Polling for Good

Instead of navigating the obstacles to conduct polls with human respondents, pollsters are running A.I. simulations instead. Why?

www.nytimes.com

oh goodie

Wow.
This is more subtle but just as alarming. Hallucinating references to scientific articles. Sure can't trust those pesky AI's!

Experiments using AI chatbots to generate papers have provided insights into how often LLMs produce citation errors and what types of error they tend to make. In one study, researchers prompted OpenAI’s GPT-4o LLM to generate six literature reviews on three mental-health disorders, and analysed the 176 references in those synthetic reviews7. Under these experimental conditions, they found that nearly 20% were fabricated references and could not be linked to actual research. And 45% of the remaining references, which corresponded to genuine publications, contained errors, often incorrect or invalid DOIs.

Hallucinated citations are polluting the scientific literature. What can be done?

Tens of thousands of publications from 2025 might include invalid references generated by AI, a Nature analysis suggests.

www.nature.com

 

[Gralin]

Wow.
This is more subtle but just as alarming. Hallucinating references to scientific articles. Sure can't trust those pesky AI's!

Experiments using AI chatbots to generate papers have provided insights into how often LLMs produce citation errors and what types of error they tend to make. In one study, researchers prompted OpenAI’s GPT-4o LLM to generate six literature reviews on three mental-health disorders, and analysed the 176 references in those synthetic reviews7. Under these experimental conditions, they found that nearly 20% were fabricated references and could not be linked to actual research. And 45% of the remaining references, which corresponded to genuine publications, contained errors, often incorrect or invalid DOIs.

Hallucinated citations are polluting the scientific literature. What can be done?

Tens of thousands of publications from 2025 might include invalid references generated by AI, a Nature analysis suggests.

www.nature.com

so AI creates a feedback loop

someone uses AI to make a paper or article or whatever

AI then picks that generated content up as part of hoovering the entire internet

and then it uses the AI generated content when creating new papers, articles etc

so as more and more incorrect info is published on the internet using LLMs the LLMs get more and more faulty info back that they give higher and higher preference to based on how much its out there

another fun one I read about yesterday was an Apple research paper that showed that none of the LLMs can count

there is a standard test they all run through for what would basically be upper primary school level maths questions

but its the same test every time

Apple changed the numbers, so the equations were the same, same level of difficulty but a different number as the answer

and there was something like an immediate 25-40% drop in accuracy

because these things are just pattern matching, they don't actually understand, its just predictive text answers

the other thing they did was with word based math questions

on friday you harvest 40 kiwi fruit
on saturday you harvest 50 kiwi fruit
on sunday you harvest double what you did on saturday
how many did you harvest over 3 days?
95% of the time they come up with 190 as the answer

change that slightly to

on friday you harvest 40 kiwi fruit
on saturday you harvest 50 kiwi fruit
on sunday you harvest double what you did on saturday but 5 of them were smaller
how many did you harvest over 3 days?
they all start changing the answer to 185

and the more convoluted you make the question the worse they do
ie the more irrelevant information you add the more they change the answer because the LLM can't parse relevance

 

[Johnny Bananas]

 

[CD Xbow]

View attachment 2622019

What about the Papal Mainframe?

 

[Chief]

Pope still hoarding wealth. And how are they doing with fighting legal battles against abused kids?

 

[Gralin]

The job loses are accelerating and the product isn't getting better

But it is getting more expensive so companies have to get rid of more staff to break even on the investment

So that's great

 

[Chief]

Have to find the speech by Zuck where he was explaining to IT staff how great it was that their AI surveillance software was watching their coding practices. Because "it will be great when we can get it to code a lot better".

 

[Not Important]

i'm with kevin.

 

[hamohawk1]

The capacity of ai will inevitably lead to some job losses, similar to every advance in technology. What is the unknown is if any jobs are created from this to shift the jobs lost over. Decision making by humans seems to be the last bastion, but from a computer that can design a new house to potentially a robot that can work non-stop to concrete, frame etc for a construcion built on ai precision it seems government and general society is flying blind.

at its best it could be complimentary to a lot of industries, at worst its genuinely society changing from what kids should be taught at school, to wtf adults do with all this time/ less money.

compare technology now to the 80's and 90's to now and it shows how we shouldnt judge what ai is now, but how progressed it could be.

 

[SLF]

The job loses are accelerating and the product isn't getting better

But it is getting more expensive so companies have to get rid of more staff to break even on the investment

So that's great

The community backlash against data centres is growing by the day too.

 

[Gralin]

The community backlash against data centres is growing by the day too.

That's because in America they do fun things like sue the town that said no so they can build the data centre anyway

 

[SLF]

That's because in America they do fun things like sue the town that said no so they can build the data centre anyway

Yeah, I don't think I'll ever quite grasp US local political arrangements.

But on the wider point, resistance to the data centres and AI in general is growing rapidly.

Wait until the bubble bursts too.

 

[Johnny Bananas]

View attachment 2622019

 

[Gralin]

Yeah, I don't think I'll ever quite grasp US local political arrangements.

But on the wider point, resistance to the data centres and AI in general is growing rapidly.

Wait until the bubble bursts too.

They pick places that can't afford the legal fight to keep them out

Then they take all their water and power

It's ****ed

 

[Chief]

They pick places that can't afford the legal fight to keep them out

Then they take all their water and power

It's ****ed

I'm waiting for the news that American politicians try to solve it by borrowing money and giving it to power and water providers to share the resources more evenly.

"Let's sell debt and put the money raised into the demand side! We'll pay it back by underfunding public services!"

Oligarchy grows even bigger with the transfer of wealth.

 

[Gralin]

I'm waiting for the news that American politicians try to solve it by borrowing money and giving it to power and water providers to share the resources more evenly.

"Let's sell debt and put the money raised into the demand side! We'll pay it back by underfunding public services!"

Oligarchy grows even bigger with the transfer of wealth.

nah they'll put the prices up and make those people pay more

 

[JHF1870]

Hello everyone,

I would like to start a discussion on the topic of AI and its impact on society, school, and work. Artificial Intelligence (AI) is rapidly evolving and transforming various aspects of our lives. AI technology has the potential to revolutionize the way we live, work, and interact with each other. However, there are concerns about the implications of AI in different areas.

One of the significant concerns is how AI is changing the job market. Some argue that AI will replace many human jobs, while others believe that AI will create new jobs and opportunities. In addition, there are concerns about the impact of AI on education. Will AI change the way we learn, and if so, how will it affect traditional teaching methods?

Another issue is the ethical implications of AI. As AI becomes more advanced, it raises questions about the ethical and moral consequences of its use. For example, how will AI impact privacy, security, and social equality? What about AI biases that can be programmed into algorithms?

So, what do you think about the impact of AI on society, school, and work? Are you optimistic about the potential benefits of AI, or do you worry about its potential drawbacks? What steps can we take to ensure that AI technology is developed and used responsibly and ethically?

Let's start a discussion and share our thoughts on this important topic.

I’m probably more optimistic than worried, but only if people stay clear-eyed about what AI actually is.

To me, AI is a tool, not a replacement for judgement. In work, school, or even analysis, it can speed things up enormously, but it can also make bad thinking look polished. That’s probably the biggest risk. Not that AI gives answers, but that people stop checking whether the answers make sense.

In education, I think the challenge is similar. If students use it to skip learning, it’s a problem. If they use it to test ideas, debug mistakes, or understand something from another angle, it could be genuinely valuable.

The same applies to jobs. Some roles will change, some tasks will disappear, and new ones will appear. But the people who can combine AI with actual domain knowledge will probably be in the strongest position.

So I’m not anti-AI at all. I just think the important bit is keeping the human part in the loop: judgement, ethics, accountability, and a bit of common sense.

For context, I have been using it to code stats with Rstudio and what started out as a blind use tool lead to a lot of frustration when I didn't understand the code, or I woke up not understanding where I was. I then used AI as a more structured tool first by drawing up a plan of what I wanted, and then making sure my layout, folder, file and naming conventions were consistent. Now I'm finding myself in more control of what I'm doing without AI doimg the heavy lifting.

I think its a mistake for people to think AI is a do it all tool. All of the ideas Ive implememted are mine and I refused to let AI guide me away from them

 

[Gralin]

I’m probably more optimistic than worried, but only if people stay clear-eyed about what AI actually is.

To me, AI is a tool, not a replacement for judgement. In work, school, or even analysis, it can speed things up enormously, but it can also make bad thinking look polished. That’s probably the biggest risk. Not that AI gives answers, but that people stop checking whether the answers make sense.

In education, I think the challenge is similar. If students use it to skip learning, it’s a problem. If they use it to test ideas, debug mistakes, or understand something from another angle, it could be genuinely valuable.

The same applies to jobs. Some roles will change, some tasks will disappear, and new ones will appear. But the people who can combine AI with actual domain knowledge will probably be in the strongest position.

So I’m not anti-AI at all. I just think the important bit is keeping the human part in the loop: judgement, ethics, accountability, and a bit of common sense.

You've fallen for the marketing if you think they want it to just be a tool

Like Sam Altman is openly calling it intelligence as a utility model

they don't want people thinking they want people paying them instead of thinking

if you know a topic and query AI you will understand how bad it is

if you don't know a topic and query AI you will think its amazing

because it is very good and sounding convincing

AI psychosis is a real issue, people have died and people have been murdered

the environmental impact is huge

the enshitification of everything is huge

using AI is not a skill, its being marketed as a skill but its not

whats the point of getting AI to do something if you then have to check everything it does to make sure its correct?

it doesn't actually save you time or make things better, and its not free even if you're not directly paying money to use it

companies are paying money to use it, and now they're getting rid of staff because they need to pay their AI bills, not because AI has increased efficiency, but because they can't afford to keep the salaries on the books

If you want to talk ethics, you can't be pro AI either

models are trained on theft, models are used to replace the people they stole the data off in the first place

so they cost people income, its a wealth trasfer to the AI company at the expense of the original creators

the data centers are polluting communities, increasing power bills for residents and taking the drinking water

there is no ethical way to consume AI models

 

[Gralin]

and yes I am aware our society makes consuming anything ethically difficult

but AI is magnitudes more damaging than most things

 

[JHF1870]

You've fallen for the marketing if you think they want it to just be a tool

Like Sam Altman is openly calling it intelligence as a utility model

they don't want people thinking they want people paying them instead of thinking

if you know a topic and query AI you will understand how bad it is

if you don't know a topic and query AI you will think its amazing

because it is very good and sounding convincing

AI psychosis is a real issue, people have died and people have been murdered

the environmental impact is huge

the enshitification of everything is huge

using AI is not a skill, its being marketed as a skill but its not

whats the point of getting AI to do something if you then have to check everything it does to make sure its correct?

it doesn't actually save you time or make things better, and its not free even if you're not directly paying money to use it

companies are paying money to use it, and now they're getting rid of staff because they need to pay their AI bills, not because AI has increased efficiency, but because they can't afford to keep the salaries on the books

If you want to talk ethics, you can't be pro AI either

models are trained on theft, models are used to replace the people they stole the data off in the first place

so they cost people income, its a wealth trasfer to the AI company at the expense of the original creators

the data centers are polluting communities, increasing power bills for residents and taking the drinking water

there is no ethical way to consume AI models

I would never have been able to do what I did with my stats without spending 6 to 12 months learning it. In the beginning I didnt understand the processes but now I have my scripts, I can do it all myself and understand the processes.

In terms of crappy ideas lauded by AI Ive been careful to understand my ideas independent of AI. AI is almost always agreeable Ive tested it with contradictory results.

im at the gym so Ill read your post in more detail when home. I don't disagree with some of the points made and appreciate your opinions.