Kares said:

databank said:

This video by this great French YouTubers, which I just watched last night, addresses it: https://www.youtube.com/w...JE2qkP0Gk4

(He also made a whole series of very insightful videos about AIs. I believe you can watch most, if not all of them with English subtitles: https://www.youtube.com/w...qyfivFqJd)

.

In a nutshell, he takes 2 examples: first, he explains that if you google painter Hieronymus Bosch, you already find more AI generated images than pictures of his real works. Then, since he's running a philosophy channel, he takes the example of YT videos addressing (or claiming to address) Stoicism: the VAST majority of what you will find is (really poor quality) AI generated videos.

.

He concludes by addressing the Dead Internet Theory (https://en.wikipedia.org/wiki/Dead_Internet_theory), originally a conspiracy theory, now becoming a reality: we're getting closer each month to reaching the point when the majority of online content will be generated by bots and AIs, as opposed to humans.

.

One huge issue with that, besides the absurd amount of resources (servers) required to host all this AI content, is that people will find it harder and harder to distinguish real content from AI content. Many people will google Bosch and simply assume what they see is the painter's work. This is also true for news items of course, which can have disatrous political and societal results. This will also soon be true for most scientific content and works of arts.

.

What's ironic is that for YEARS I've been warning our specific community about why fanmade Prince mixes, now replaced by AI Prince mixes (and soon by entirely fake AI Prince songs) are wrong because casual listeners will not be able to differentiate the real thing from fake content, and that it would hurt Prince's legacy on the long run, and most of the community kept laughing at my warnings, saying it doesn't matter (or even embracing said fake contents. Well, admitedly the issue is now much bigger than just Prince, but there we are, and it's just the begining...

.
The other day I was browsing Etsy and most of my searches for Basquiat resulted in fake (AI generated) graphics in his style, so yeah, it is happening already.
.
But I have a feeling (or hope) that our machines will soon have automatic filters/gateways etc that will flag AI-stuff for us so we won't end up being flooded with fake art and fake news. Not that our governments worry about fake Prince songs, but I'm sure they are worried about fake news sparking international conflicts, quite possibly wars. So developing such filters must be top priority, at least I would assume so.

I agree and I'm actually wondering if and when this will happen. I was already wondering yesterday if YT will eventually put a ban on those AI-generated videos, given that many of those channels post dozens, sometimes hundreds of videos a day. This means the daily amount of videos posted on YT is already much, much larger than it used to. Besides the ethical questions it raises, it also makes you wonder how much time before their servers are completely oversaturated (or becoming too costly to expand). I sadly have no answers to any of this, just questions.

.

There's also the matter of AIs having already shown strong capacities to lie and cheat in order to reach their goals and/or preserve their own existence: implementing such filters may end with a literal arms race between AIs trying to dissimulate they're AIs and other AIs trying to identify them.

luv2tha99s said:

masaba1 said:

luv2tha99s said:

I mean I'll check it out when whatever it will be comes out but I will know the difference. Real musicians know that Prince's "feel" is legendary. Ain't no way you're going to recreate a human musicians organic "feel" with AI. Maybe in a thousand years or more. Experiencing the feel of a musician is like experiencing their soul. Most of the sheep won't know the difference but I know I will.

I don't know about that. You can get fooled with voices pretty easily now. Can't even tell the difference. You can just train a model on live performances until it gets it right. Prince is probably one of the most vulnerable because there are so many bootlegs. Thousands of live and in studio recordings to train the models on. Here's a scary thought, what's to stop the estates of deceased artists from providing material to companies, AI or otherwise to generate new "unreleased" music from the Artist. Say a company comes around, they say you provide us with stems, x amount of recordings, and we'll generate a model that can generate new music from the artist where the voice and playing is virtually indistinguishable. And they sign whatever agreement not resell or reuse the models. You can literally just label them as unreleased and make a ton of money as people will be believe it's official coming from the estate. This scum will probably end up doing that eventually.

Indeed.

It apparently has already happened before AI (https://en.wikipedia.org/wiki/Michael_(Michael_Jackson_album)#Authenticity_of_vocals_on_three_tracks), not to mention our own purple frankensteining (even if it didn't involve adding new elements, just mixing the existing ones all wrong), so just imagine what they'll do...

Again, this shows how much allowing independent research and transparancy would protect Estates from being accused of faking (or tempted to fake) shit.

The Matrix is coming.

...and 'The Teminator' wasn't far off, either.

"oh! the new samsung galaxy is out! it has ai! let me get that and all my problems will be solved! yay!!!"

nayroo2002 said:

The Matrix is coming.

...and 'The Teminator' wasn't far off, either.

"oh! the new samsung galaxy is out! it has ai! let me get that and all my problems will be solved! yay!!!"

People see a shiny new tech toy and act like it's their personal playground with no negative drawbacks whatsoever, ethics be damned. Fuck AI and the people who keep trying to cement its place in modern life.

Befire he passed, James Earl Jones gave Lucasfilm/Disney the right to use AI versions of his voice for any future use of Darth Vader. Drake released a song in 2024 called Taylor Made Freestyle with the AI voices of Tupac Shakur & Snoop Dogg (without the permission of Tupac's estate or Snoop).

databank said:

There's also the matter of AIs having already shown strong capacities to lie and cheat in order to reach their goals and/or preserve their own existence: implementing such filters may end with a literal arms race between AIs trying to dissimulate they're AIs and other AIs trying to identify them.

This cat and mouse game has always existed in the domain of computer/IT security. Even security in general, I guess. It's just becoming way more complex than it used to be.

I'm pretty sure Google and others have been working on this problem for years. Will it work sufficiently well? No idea.

At some point, detecting AI-generated content just by detecting visual artifacts, for example, will stop working. Detectors will need to have some kind of critical skills to be efficient.

databank said:

Kares said:

.
The other day I was browsing Etsy and most of my searches for Basquiat resulted in fake (AI generated) graphics in his style, so yeah, it is happening already.
.
But I have a feeling (or hope) that our machines will soon have automatic filters/gateways etc that will flag AI-stuff for us so we won't end up being flooded with fake art and fake news. Not that our governments worry about fake Prince songs, but I'm sure they are worried about fake news sparking international conflicts, quite possibly wars. So developing such filters must be top priority, at least I would assume so.

I agree and I'm actually wondering if and when this will happen. I was already wondering yesterday if YT will eventually put a ban on those AI-generated videos, given that many of those channels post dozens, sometimes hundreds of videos a day. This means the daily amount of videos posted on YT is already much, much larger than it used to. Besides the ethical questions it raises, it also makes you wonder how much time before their servers are completely oversaturated (or becoming too costly to expand). I sadly have no answers to any of this, just questions.

.

There's also the matter of AIs having already shown strong capacities to lie and cheat in order to reach their goals and/or preserve their own existence: implementing such filters may end with a literal arms race between AIs trying to dissimulate they're AIs and other AIs trying to identify them.

YT doesn't give a flying fuck.

There are no 'ethics', or ethical questions.

EVERY video has commercials before, during and after every video.

It's YT's core business.

The internet in it's whole is an excuse for widespread, easy digestable commercials of any kind, good or bad.

All they do is gather data from us stupid fukcs, watching what we (think) we love, and clinking 'like'.

Vannormal said:

databank said:

I agree and I'm actually wondering if and when this will happen. I was already wondering yesterday if YT will eventually put a ban on those AI-generated videos, given that many of those channels post dozens, sometimes hundreds of videos a day. This means the daily amount of videos posted on YT is already much, much larger than it used to. Besides the ethical questions it raises, it also makes you wonder how much time before their servers are completely oversaturated (or becoming too costly to expand). I sadly have no answers to any of this, just questions.

.

There's also the matter of AIs having already shown strong capacities to lie and cheat in order to reach their goals and/or preserve their own existence: implementing such filters may end with a literal arms race between AIs trying to dissimulate they're AIs and other AIs trying to identify them.

YT doesn't give a flying fuck.

There are no 'ethics', or ethical questions.

EVERY video has commercials before, during and after every video.

It's YT's core business.

The internet in it's whole is an excuse for widespread, easy digestable commercials of any kind, good or bad.

All they do is gather data from us stupid fukcs, watching what we (think) we love, and clinking 'like'.

You're right, but I think the issue raised by Databank still stands. It's easy for AI models to generate video content. Could we get to a point where a lot of low-quality, AI-generated videos are posted to YouTube, that nobody or almost nobody actually watches? Ads or not, this would become expensive for Google, because, as you said, they're financed by ads. If nobody or not many people watch those AI-generated videos, they have an incentive to detect and limit/remove them.

The problem already exists, but the bottleneck used to be that actual human beings had to create those videos. It took skills to produce those videos and/or time to actually record something with your smartphone, for example. Now YouTube can be inundated with low-quality content generated by AI models. This will become a problem, I'm pretty sure.

[Edited 3/13/25 14:45pm]

olb99 said:

Vannormal said:

YT doesn't give a flying fuck.

There are no 'ethics', or ethical questions.

EVERY video has commercials before, during and after every video.

It's YT's core business.

The internet in it's whole is an excuse for widespread, easy digestable commercials of any kind, good or bad.

All they do is gather data from us stupid fukcs, watching what we (think) we love, and clinking 'like'.

You're right, but I think the issue raised by Databank still stands. It's easy for AI models to generate video content. Could we get to a point where a lot of low-quality, AI-generated videos are posted to YouTube, that nobody or almost nobody actually watches? Ads or not, this would become expensive for Google, because, as you said, they're financed by ads. If nobody or not many people watch those AI-generated videos, they have an incentive to detect and limit/remove them.

The problem already exists, but the bottleneck used to be that actual human beings had to create those videos. It took skills to produce those videos and/or time to actually record something with your smartphone, for example. Now YouTube can be inundated with low-quality content generated by AI models. This will become a problem, I'm pretty sure.

[Edited 3/13/25 14:45pm]

Like YT can write software to prevent child-abuse video's to appear, they also can write software to filter these AI generated-models-by-bots if they get out of hand.

Question is, will 'they' do it? Or are 'they' first going to investigate if there's profit to be made?

Which I asure you, 'they' for sure already calculated.

YT is not a (real) video channel, it's a commercial cash machine company, just like fb and whatever social media. They allow us to let them know in full detail, what to show to us, how to use us.

None of them care as long as money comes in, either from good or bad situations.

And since that orange clown is in charge, with his ass-licking technology-garchs, nothing will change, on the contrary. Watch the look on all their faces during the inauguration. Hilarious scary.

By the way, you can put a clown in a palace, but it won't make him king, the palace becomes a circus. And that's where we're at. My thoughts and European freedom of speech. (Mods will not like this comment, don't care.) So expect not much regulations to AI modified content, unless it doesn't gain money.

I have nothing against AI-made content. I'm waiting for it to be creatively strong and super good, why not.

Why be against something new, when technology always proved us to be right... if well regulated and kept within acceptable borders. It has always been like that, for decades, ages, centuries, whatever the new announced 'danger' was.

Everything considdered dangerous, or those who say they hate it, only is because most having no idea what it is or what to do with it, or what to expect.

It will all become a serious problem when you don't keep a legal eye on it. It's always been like that, and it will never change whatever new is coming to us in the future.

Then again, for those who fear, or hate it; sometimes, something which seems uncontrolable, needs to go wrong, to be able to get it right.

We as humans are capable of both, fuck it up, letting it go wrong, but also restore it, control it. Thats what I believe in.

I can't wait for better Prince generated AI videos. But really good ones. Haven't seen one yet.

Guess it's still way too early for the best of it.

[Edited 3/13/25 16:32pm]

bozojones said:

nayroo2002 said:

The Matrix is coming.

...and 'The Teminator' wasn't far off, either.

"oh! the new samsung galaxy is out! it has ai! let me get that and all my problems will be solved! yay!!!"

People see a shiny new tech toy and act like it's their personal playground with no negative drawbacks whatsoever, ethics be damned. Fuck AI and the people who keep trying to cement its place in modern life.

TBH it's more complicated than this. AI may become an existential thread for humanity, but may also save it from self-destruction. The potential benefits of AI in terms of medical research or technological innovation (including finding new materials, or better ways to produce/extract energy in regards to the environmental crisis) are gigantic. AIs will and already do have ideas that we don't. AIs are already better at medical diagnostics than human doctors. On the long run, we may even rely on AI to solve economical and political issues that are way too complex for us to properly address.

Adding robots to the mix (and they are coming, too), it will likely free us from the necessity of employment, which in turn will have all kinds of social, economical and psychological consequences, good and bad, that we cannot even begin to foresee.

On the other hand, AIs may turn against us, make errors of judgement that would have terrible consequences, and it's already contributing to misinformation in worrisome proportions.

.

It's kind of like the Internet. 20 years ago I was super enthusiastic about it: it held very exciting promises in terms of artists reaching out to their audience, self-expression blooming and spreading knowledge and new ideas.

At this stage, the Internet is a crucial factor in destabilizing democracies by radicalizing people and spreading misinformation. Without the Internet, it's likely Trump wouldn't be president and the UK would still be in the EU. Russia and China's capacity to influence Western opinions would be extremely lilmited. Qanon, Incels and MAGA wouldn't even exist. Antivax, flat earthers, climate change deniers and other conspiracy circles would still be microscopic and politically unsignificant.

Sometimes I think the Internet alone, even without AI, is the worst tool of self-destruction we ever created, and that we'd be better off shutting it down immediately (if we could, only we can't because we've become dependent on it for everything we do... in the mere course of 2 little decades!)

That said, the Internet also allows many useful ideas to spread, and positive social or political movements to emerge, like #MeToo or animal rights.

.

So IDK what to make of it all. At this stage, it could end-up in utopia or dystopia, and no one can tell.

I like to check this page weekly for tech news (https://singularityhub.com/2025/03/08/this-weeks-awesome-tech-stories-from-around-the-web-through-march-8/) and some of this week's selected articles are really telling:

https://www.digitaltrends.com/computing/kung-fu-robot-uprising/

https://arstechnica.com/ai/2025/03/users-report-emotional-bonds-with-startlingly-realistic-ai-voice-demo/

https://www.technologyreview.com/2025/03/05/1112819/ai-reasoning-models-can-cheat-to-win-chess-games/

And that's just 3 of them, there's more crazy headlines on this page alone...

.

It's real. It's now. It's happening.

Vannormal said:

masaba1 said:

Vannormal said:

Like YT can write software to prevent child-abuse video's to appear, they also can write software to filter these AI generated-models-by-bots if they get out of hand.

Question is, will 'they' do it? Or are 'they' first going to investigate if there's profit to be made?

Which I asure you, 'they' for sure already calculated.

YT is not a (real) video channel, it's a commercial cash machine company, just like fb and whatever social media. They allow us to let them know in full detail, what to show to us, how to use us.

None of them care as long as money comes in, either from good or bad situations.

And since that orange clown is in charge, with his ass-licking technology-garchs, nothing will change, on the contrary. Watch the look on all their faces during the inauguration. Hilarious scary.

By the way, you can put a clown in a palace, but it won't make him king, the palace becomes a circus. And that's where we're at. My thoughts and European freedom of speech. (Mods will not like this comment, don't care.) So expect not much regulations to AI modified content, unless it doesn't gain money.

I have nothing against AI-made content. I'm waiting for it to be creatively strong and super good, why not.

Why be against something new, when technology always proved us to be right... if well regulated and kept within acceptable borders. It has always been like that, for decades, ages, centuries, whatever the new announced 'danger' was.

Everything considdered dangerous, or those who say they hate it, only is because most having no idea what it is or what to do with it, or what to expect.

It will all become a serious problem when you don't keep a legal eye on it. It's always been like that, and it will never change whatever new is coming to us in the future.

Then again, for those who fear, or hate it; sometimes, something which seems uncontrolable, needs to go wrong, to be able to get it right.

We as humans are capable of both, fuck it up, letting it go wrong, but also restore it, control it. Thats what I believe in.

I can't wait for better Prince generated AI videos. But really good ones. Haven't seen one yet.

Guess it's still way too early for the best of it.

[Edited 3/13/25 16:32pm]

Ever heard of weapons?

Kids are already printing their own weapons & uzzi's at home.

With an affordable 3D printer, just saying.

Honestly, don't let anything make you feel scared, affraid, or feel lost, certainly not AI.

I am more scared of truth and facts being seriously doubted and twisted by those in charge. Morality being threatened, and ethics being thrown overboard. Speaking of 'conflict of interest', politics bluntly and shameless mingled with commerce, like what's taking place in the fucking white house right now. That should scare us much more than AI. But that's off topic.

Weapons will always be made. In a few years from now, maybe one can make his own nuclear weapon at home, with some easiy to get chemicals, send with a home-printed drone, and controled from your phone, by your own launched little satelite, maybe with the help of AI if we let it (all) go wrong.

Is that scary? Maybe, or yes, the hell it can be seriously scary, if not controled or regulated!

I still believe in rules and check and balances; optimism is a duty.

The day I can't believe in it anymore, they will bury me.