Master AI (Artificial Intelligence) Discussion/News Thread

[sonitus mirus]

Here's a video I found interesting and which exposes a problem with AI I can relate to, as many of you I guess:

Yes, sometimes I have to apply clear, simplified logic to "rationalize" why something stated cannot be true, or is at least contradictory, before the AI will capitulate and admit there was an error in their logic. Using a paid version of ChatGPT with, supposedly, the most advanced models.

 

[Snarfie]

How to absord information these days is almost a mental effort. As boedism says you are what you think but you are not your thoughts.

 

[Puddingbuks]

 

[DayDiner1969]

My mum showed me a video on Facebook earlier which purported to show Madeleine McCann (3yr old who went missing in Portugal 18years ago) being released from an concealed underground bunker in Portugal, it had thousands of likes/shares and many comments underneath. The ability of gullible folk to believe everything they see/hear on such gutter shit media sites has changed the world we live in for the worse. AI is not going to be the savour of the world, it will only fuck those who are not the 1%ers

Madeleine McCann search in Portugal ends after three days

Authorities carried out a fresh search near where the three-year-old vanished in Praia da Luz in 2007.

www.bbc.co.uk

If I had my way you should need to provide a real name/ evidence of who you are before you get the ability to sign up for such platforms, this would obviously impact whistleblowers/political dissidents etc so some form of protection would need to be in place to maintain their anonymity but I don't have the brains to work out a solution for such scenarios

We certainly don't seem to need any AI for this to be a huge problem. The future seemed a lot less bleak before they made it easy enough for all the stupid people to get online. And I am not talking about "people who disagree with me" or "people who don't know calculus" or anything like that. I mean the people who are incredibly confident that the lizard people are injecting 5G into our chemtrails to control our minds. And even then, it would be mostly tolerable if they didn't insist on being in charge of the government. I'm not sure we'll get through this one without some pain.

 

[Esprit]

https://ml-site.cdn-apple.com/papers/the-illusion-of-thinking.pdf

 

[Sir Sanders Zingmore]

The ability of gullible folk to believe everything they see/hear on such gutter shit media sites has changed the world we live in for the worse.

The thing that worries me is that soon (possibly sooner than we think), it will no longer be possible for *anyone* to tell the difference between real and fabricated content.
Where does it leave us if we cannot trust anything we see online?

 

[Karmacoma]

The thing that worries me is that soon (possibly sooner than we think), it will no longer be possible for *anyone* to tell the difference between real and fabricated content.
Where does it leave us if we cannot trust anything we see online?

We'll have an AI for that. And it'll be just like the rest, we'll have to rely on our state structure and in democracy (the modern definition of democracy) as a whole to give us such tools.

 

[Puddingbuks]

The thing that worries me is that soon (possibly sooner than we think), it will no longer be possible for *anyone* to tell the difference between real and fabricated content.
Where does it leave us if we cannot trust anything we see online?

Online you can do your own research to some degree. It is clear to me that many news outlets are heavily biased in their reporting. Some news or details just will be ignored.

I only have to read the headline and who it is about and I know what is coming.

But I agree that it will be difficult if videos or people talking are not real.

 

[Sir Sanders Zingmore]

We'll have an AI for that. And it'll be just like the rest, we'll have to rely on our state structure and in democracy (the modern definition of democracy) as a whole to give us such tools.

I expect that soon (sooner than we think) we will get AI content that can't be distinguished from real content even by other AI

But I agree that it will be difficult if videos or people talking are not real.

This is where I think it gets scary. Hypothetical scenario: video is leaked of the CEO of a prominent public company snorting illicit drugs. CEO denies it saying it's AI generated. There'll be literally no way of knowing if they are telling the truth.

 

[Chr1]

Aye. Covert counterinteligence, political influence via social media undermining companies, democracies etc. Already a thing. It will just be a lot easier for anyone to do. And harder/impossible to distinguish from reality.
Tip of the iceberg unfortunately.

 

[bluefuzz]

it will no longer be possible for *anyone* to tell the difference between real and fabricated content.
Where does it leave us if we cannot trust anything we see online?

Has it ever been possible? Could we ever trust the printed word, paint on canvas, or a stone inscription? Humans have, as long as they have been able to call themselves 'human', used the latest technologies to spread lies, fakery and disinformation to further their religious and/or political agendas while feathering their own nests. And people have largely been gullible enough to fall for most of it. It's not a coincidence that the spread of the major world religions began not long after the invention of writing itself. And there are still billions of people in the 21st century that believe in the stories of some stone-age middle-eastern and asian goat-herdsmen simply because they happened to have the bright idea of committing their fantasies to clay-tablet, papyrus and parchment.

The 'problem' with generative AI is not so much the technology itself, but who gets to set the agenda of how, where and why it gets used. Just as it has been for other forms of disruptive technology such as the wheel, the stirrup, the reed pen, moveable type, the steam engine, radio, television etc. ad infinitum ...

 

[Esprit]

"The man who reads nothing at all is better educated than he who reads nothing but newspapers."
Thomas Jefferson

 

[Sir Sanders Zingmore]

The 'problem' with generative AI is not so much the technology itself, but who gets to set the agenda of how, where and why it gets used. Just as it has been for other forms of disruptive technology such as the wheel, the stirrup, the reed pen, moveable type, the steam engine, radio, television etc. ad infinitum ...

I highly recommend this book

'NEXUS' – A new book from Yuval Noah Harari (out September 2024)

'NEXUS: A Brief History of Information Networks from the Stone Age to AI' (published Sept 2024) is a non-fiction book by Yuval Noah Harari. In this thrilling new work, the historian and bestselling author looks through the long lens of human history to consider how the flow of information has...

www.ynharari.com

Which argues that the technology itself is the problem because it is completely unlike other forms of disruptive technology. It can make decisions and create ideas by itself. In other words the "who"who sets the agenda is the technology itself rather than say a human spreading human ideas by controlling a printing press

 

[OCA]

Has it ever been possible? Could we ever trust the printed word, paint on canvas, or a stone inscription? Humans have, as long as they have been able to call themselves 'human', used the latest technologies to spread lies, fakery and disinformation to further their religious and/or political agendas while feathering their own nests. And people have largely been gullible enough to fall for most of it. It's not a coincidence that the spread of the major world religions began not long after the invention of writing itself. And there are still billions of people in the 21st century that believe in the stories of some stone-age middle-eastern and asian goat-herdsmen simply because they happened to have the bright idea of committing their fantasies to clay-tablet, papyrus and parchment.

The 'problem' with generative AI is not so much the technology itself, but who gets to set the agenda of how, where and why it gets used. Just as it has been for other forms of disruptive technology such as the wheel, the stirrup, the reed pen, moveable type, the steam engine, radio, television etc. ad infinitum ...

False narrative is central to human cohesion.

 

[Sir Sanders Zingmore]

False narrative is central to human cohesion.

Have you read "Nexus"? He (Harari) talks about narratives and human cohesion at some length (actually that's a theme running through most of his books)

 

[OCA]

Have you read "Nexus"? He (Harari) talks about narratives and human cohesion at some length (actually that's a theme running through most of his books)

I haven't read Nexus yet but read Sapiens and Homo deux.

 

[bluefuzz]

I highly recommend this book

I wasn't much impressed by what I've previously read by Harari (Sapiens etc.). He seems to me to be a pupulist scaremonger whose claims rarely stand up to close scrutiny. A lot of hand waving and what if'ery.

it is completely unlike other forms of disruptive technology. It can make decisions and create ideas by itself.

I've yet to see any evidence that either of those claims are, or will be, true anytime within the next century or two ...

 

[Sir Sanders Zingmore]

I've yet to see any evidence that either of those claims are, or will be, true anytime within the next century or two ...

Already happening. AI already decides who gets interviewed when applying for jobs in some companies. It decides what you see in your newsfeeds (take that Gutenberg) and it advises on court sentencing. There's a fully AI virtual hospital in China. An enormous percentage of all trades in financial markets happen with zero human intervention.

 

[OCA]

Already happening. AI already decides who gets interviewed when applying for jobs in some companies. It decides what you see in your newsfeeds (take that Gutenberg) and it advises on court sentencing. There's a fully AI virtual hospital in China. An enormous percentage of all trades in financial markets happen with zero human intervention.

If you're regularly using a keyboard in your job, you will most likely be jobless within the next 5 years.

 

[Mort]

If you're regularly using a keyboard in your job, you will most likely be jobless within the next 5 years.

I'll take the over on that.