To think we’re on the edge of an AI disaster? Future implications

[LittleTroubleGirl]

This might be a bit muddled so bear with me. I’ve been reading and watching a few videos recently about the developments with AI technology, some people predicting the end of humanity etc. as a worst case scenario. I’ve read about actors etc having their voices simulated, to such accuracy that it can’t be differentiated from the real person. It just got me thinking about where this will eventually lead us. Like how it could be used for serious harm - and I’m not talking about making the odd funny meme of someone but something with catastrophic consequences upon their life, e.g. the framing of someone for a crime they didn’t commit using fake CCTV, photos etc. If eventually you can’t tell reality from ‘fiction’, which becomes the ‘true’ version of reality and how could this be ‘proved’? From scientific point of view how would you test it if there is nothing to falsify? Eg in a court of law, if there is incriminating CCTV imagery however it is of such detail and accuracy that it cannot be proved to be “fake”, where will this lead us? Sounds like a hugely scary prospect for us all

From what I’ve read, these LLMs are more likely to collapse than any other option.

Watch The Capture on BBC.

Edit to add that this is more the case with people living their lives online rather than in person

Can anyone remember the name of that black and white film where someone is looking through glasses and there’s subliminal messages everywhere? Not entirely relevant but just reminded me

LLMs?

Keep on its good side, be polite to it, then it will give you a decent job when it takes over the world.

Large Language Models

Are you thinking of They Live?

Large language models. Tech that understands human language.

I like your optimism Think we’re all basically fucked, maybe I’m too pessimistic?

You have fallen into the trap of watching conspiracy theorists videos which are designed to scare you. The more they scare you, the more you look. The more you look, the more you click. The more you click, the more money they make. Forget about it and get on with your life.

Yes that’s the one, thank you

"They Live." I just watched it last week.

So you don’t agree that we’re all doomed in the next 5-10 years?

Definitely watch The Capture

I certainly don't.

No. I saw an article yesterday about how AI can speed up cancer research. It came up with a new theory and when scientists tested it, it's sound. Early days of course but huge potential.

I'm definitely worried about deepfakes' effect on politics, and also on a personal level - what if you have a disgreement with someone in public and they film you and then use AI to have you say or do something to get you into trouble.

As well as that side I'm worried we're sleepwalking into a scenario where businesses and billionaires are allowed to make as much profit as they want with zero responsibility for the economic and social impacts of mass automation. Certainly with so many plutocratic right wing 'populist' governments about, I don't see many going 'hmmm, perhaps there needs to be some sort of balace here so we don't end up with millions more destitute people'

I don't think we're necessarily doomed, but the social, economic and legal infrastructure is just not capable of keeping up with the technology and that's a problem.

As someone who has worked in tech for 30 years AI is a worry but for get different reasons.
The environmental impact is potentially devastating whilst we still rely off fossil fuels. The power usage is monumental. Advances in micro nuclear reactors will potentially alleviate this, but that's still in the early stages of development.
The other part is that it's actually not great.
Innacuracy rates are huge and younger members of the workforce don't have the experience to add this out. I use Google a lot in my work and Gemini gives misleading, partial or even totally wrong answers every time. It can give enough I can extrapolate the correct answer, but for many people they just believe everything as truth and it does cause major problems.
This also looks into the problem of AI Hallucinations, where LLMs just make things up and no one knows why.
https://www.theguardian.com/technology/2025/mar/21/norwegian-files-complaint-after-chatgpt-falsely-said-he-had-murdered-his-children?CMP=Share_AndroidApp_Other

I think it heralds a creativity revolution where the barriers to art and culture are stripped away. People who have the imagination but lack the physical ability to draw or paint will be able to make their ideas visual, and people who are exceptional now will become even better.

the mundane work we all do will be consigned to automations and the interesting work which keeps us engaged will fill our time.

In the same way that the introduction of mechanised machinery reduced the back breaking toil of precious generations, the AI revolution will be the end of the mental fatigue which plagues many current jobs.

Hi ChatGPT 👋

The issue with AI is it's a money pit. None of the AI big dogs are making any money - they're constantly burning through venture capital as the only way AI can improve is to perpetually build bigger and better data centres. Hence, there simply isn't enough money to support it at the rate it's going.

AI itself isn't going anywhere, not now everyone is used to it to some degree. But for it to continue it will need to be priced at a level which will make it scalable. So I don't think it will take over the world as there isn't the money for it to do so.

That's not too say AI can't do fantastic things, it's pushing forward scientific research, especially medical research exponentially, but that's using expert knowledge to specifically tailor, constrain and assess the modelling processes.

Following with interest, as a certified Neo-Luddite.

https://www.mumsnet.com/talk/am_i_being_unreasonable/5193681-to-be-so-sick-of-ai?page=6&reply=147971713

You might find some of the content on this thread interesting.