AI art and creative content creation

Status
Not open for further replies.
I dare not post any of the images, but if you want some high quality hilari-trolling, look up #dignifAI on Twitter. in short, people are using AI to *add* clothes to images of scantily clad women. Many of the results are as flawed as we've come to expect from AI generated human images, but many look respectable... unlike the originals.

It seems a lot of people are annoyed by this.
 
people are using AI to *add* clothes to images of scantily clad women
Wonder if that qualifies as one of those "for every action there is an equal and opposite reaction" things from back in science class?

Went playing in Google to find articles about the thing.
This comment on one is of interest,

lolzers says
February 5, 2024 at 6:22 pm
They’re mad when their clothes are AI removed, they’re mad when their clothes are AI added. They’re just plain mad. ____ mad. lol

While the commenter may or may not grasp the concept, that is exactly the issue.
AI is being employed as a tool to create a presentation of them, an apparent reality of them, which is not by their will or their choosing.
That is a form of boundary crossing.

How far will that go and for how many people will it go?

For instance ...

My creative writer brain comes up with, What happens when a coworker with an axe to grind sends both your employer and the police half a dozen photos of you wearing a t-shirt with printing calling for violence against your employer's family?
Okay, you somehow conclusively prove you did no such thing and have no such attitude - what emotions about the matter now continue lingering on in your employer??? What law enforcement database are you now forever in???
 
Last edited:
My creative writer brain comes up with, What happens when a coworker with an axe to grind sends both your employer and the police half a dozen photos of you wearing a t-shirt with printing calling for violence against your employer's family?

This not new, either in fiction or reality. There are *numerous* examples of "fake tweets" and "fake texts." How do you prove you *didn't* send out those career-endings tweets where you advocated genocide or proclaimed your unhealthy love of the underaged or the most recent episodes of Doctor Who? With enough effort - and expense - you can doubtless conclusively show that the screenshots of those messages are in fact fake. But by then the reputational damage is done and your home has been SWATted and your employment terminated and your favorite hangout set ablaze.

The internet has *long* been filled with celebrity heads photoshopped atop the bodies of others in advanced states of undress. Now it can be done virtually instantly via AI, with no original photo for someone to find and prove the image fake. Look for photos and videos of political candidates doing and saying things they didn't. And look for political candidates actually caught on video saying unfortunate things claiming those videos are AI-fakes. We are at the end of photos and videos being actual evidence.

Never mind the politics... here's a report on a campaign video put out by pro-Trump people that has an AI-faked audio of Paul Harvey going on about Trump. If you know Paul Harveys voice, this is a *pretty* accurate and convincing piece, until you realize the timing doesn't make a bit of sense given that he died 15 years ago. It's cringey, but it's an indicator of what's coming.

View: https://www.youtube.com/watch?v=lIYQfyA_1Hc
 
My creative writer brain comes up with, What happens when a coworker with an axe to grind sends both your employer and the police half a dozen photos of you wearing a t-shirt with printing calling for violence against your employer's family?
By now it's so common, that actually the opposite problem is more likely - that police would ignore real threats, assumung them to be just another dumb joke.
 
I wonder why on most of the AI image generators I've tried, it's impossible to generate a jet, or a realistic aircraft carrier, while the AI succeeds very well in drawing tanks ?
17072985016982pwyk4n6 - Copie.png 17073170974961iq4alpb.png

1707383479489cd0f671y.png 170738346726762836drm.png
 
Lack of data and lack of effort imo. A huge amount of work has been put into generating good human images both in data collection and model hacks since that is where the demand is at. I've seen some decent aircraft generations with tuned models, but haven't seen a model that understands what exactly is an aircraft carrier and what it should look like.
 
Lack of data and lack of effort imo. A huge amount of work has been put into generating good human images both in data collection and model hacks since that is where the demand is at. I've seen some decent aircraft generations with tuned models, but haven't seen a model that understands what exactly is an aircraft carrier and what it should look like.
"Artificial Intelligence" (and I use that term very advisedly) at this stage is literally and figuratively still (thankfully) *mindless*, so any results it produces are pretty much based on online popularity or prevalence, rather than on any self-directed, self-evolved agency. Once again, AI will only truly become a Menace to Society if it somehow also acquires Artificial Consciousness along the way. Considering that this feat appears to still be a tall order for any good (bad?) number of mouth breathing humans, I'm not in any way inclined to hold my nose breath on that one...
 
Last edited:
So far, one of the "most realistic" aircraft carriers I've had generated.
"Vintage aircraft carrier"
vintage_aircraft_carrier_seed3031602321132422_029.jpg
 
Last edited:
So far, one of the most realistic aircraft carriers I've had generated.
"Vintage aircraft carrier"
View attachment 719497
Looks vaguely like some Soviet 1930s brutalist propaganda poster art (minus the US flag, of course). But if you're into that kind of kinky stuff...
 
Last edited:
If you know Paul Harveys voice, this is a *pretty* accurate.... It's cringey
His announcer was worse...it always sounded to me that he said:

"And now, Paul Harvey nude."

AI info
 
Last edited:
February 15, 2024
AI attempts to draw British aircraft and we spit out our collective tea in awe at these magnificent obscenities

(Brief note on method: the aircraft name and nothing else was typed into two freely available image generation websites, Gencraft and Da Vinci. These are the genuine results).

 
So that's what mindless intelligence really looks like - we sure still have a long way to go until we can welcome our AI overlords...
 
Last edited:
So that's what mindless intelligence really looks like - we sure still have a long way to go until we can welcome our AI overlords...

Elon Musk is waiting for the brain to insert into his Teslabot. As well as the plasma-pulse rifle...
 
Are you really sure about the intelligence part though?
Yes. Many are quite intelligent. A few are of course monumentally stupid, getting to their positions of power purely because they've been dragged there by their puppetmasters... but most are filled with a supreme dark intelligence. It is an intelligence devoid of humanity, a political psychopathy. They can calculate a strategy. But they cannot conceive of a theory of mind to understand those who might disagree with them. The do not empathize... they just learn to pretend to.

Those who argue against the future success of AI in various endeavors fail to realize that human NPCs seem to be real. A great many humans get along just fine without a lot of what many of us think makes humanity great. AI need not replicate the best of us to succeed. It just needs to replicate a *sufficient* human. A sociopathic AI-artist is not a problem, of course. A sociopathic AI in charge of things that actually matter, though, could well be an issue.
 
Yes. Many are quite intelligent. A few are of course monumentally stupid, getting to their positions of power purely because they've been dragged there by their puppetmasters... but most are filled with a supreme dark intelligence. It is an intelligence devoid of humanity, a political psychopathy. They can calculate a strategy. But they cannot conceive of a theory of mind to understand those who might disagree with them. The do not empathize... they just learn to pretend to.

Those who argue against the future success of AI in various endeavors fail to realize that human NPCs seem to be real. A great many humans get along just fine without a lot of what many of us think makes humanity great. AI need not replicate the best of us to succeed. It just needs to replicate a *sufficient* human. A sociopathic AI-artist is not a problem, of course. A sociopathic AI in charge of things that actually matter, though, could well be an issue.

With all due respect, I don't consider "achieving" human mediocrity to be a worthy goal. As my employer recently said about storytelling as regards writers: "I want heroic characters who are heroic."
 
An attempt to create a 'Chinese Supercarrier' using AI that appeared as a thumbnail on a military clickbait channel. And that's where I see most of the products of this kind of 'Procedural Generation' as thumbnails on YouTube.
 

Attachments

  • Military_Clickbait (Chinese Supercarrier).png
    Military_Clickbait (Chinese Supercarrier).png
    126 KB · Views: 16
With all due respect, I don't consider "achieving" human mediocrity to be a worthy goal.
For a *machine?* Human mediocrity would be *spectacular.* Let's say, instead of something trivial like art, we had AI who turned into mediocre, average human doctors and diagnosticians. Would you rather have a *great* human diagnostician than a mediocre AI one? Sure. But you could actually *have* a mediocre AI diagnostician in your home. It non invasively scans and examines you three times a day. In contrast, one American a week gets to see Doctor House, and then pay out the ear for the privilege.

Similarly, the police could have ten thousand mediocre detectives scanning the evidence nonstop, where the good human detectives can work on one case at a time. A million mediocre AI astronomers scanning the skies looking for comets, asteroids and aliens.

An army of mediocres could do vastly more good than a few experts. And these are mediocres that simply didn't exist until humans invented them.


As my employer recently said about storytelling as regards writers: "I want heroic characters who are heroic."
Yeah, but those aren't real people, and thus the audience won;t connect with them. Or so I've been repeatedly told.
 
Are these perfect? Nope. A spoon magically appears in the old womans hand; the woman walking down the street in Japan leaves no footprints. But they're *close.* For snippets of "stock footage," these are just about there. The Woolly Mammoth video could be used in a cable documentary about many things. For background videos - stuff barely seen on screens behind the main characters - this sort of thing would seem fantastically useful.
 
For a *machine?* Human mediocrity would be *spectacular.* Let's say, instead of something trivial like art, we had AI who turned into mediocre, average human doctors and diagnosticians. Would you rather have a *great* human diagnostician than a mediocre AI one? Sure. But you could actually *have* a mediocre AI diagnostician in your home. It non invasively scans and examines you three times a day. In contrast, one American a week gets to see Doctor House, and then pay out the ear for the privilege.

Similarly, the police could have ten thousand mediocre detectives scanning the evidence nonstop, where the good human detectives can work on one case at a time. A million mediocre AI astronomers scanning the skies looking for comets, asteroids and aliens.

An army of mediocres could do vastly more good than a few experts. And these are mediocres that simply didn't exist until humans invented them.



Yeah, but those aren't real people, and thus the audience won;t connect with them. Or so I've been repeatedly told.

Looking through the Internet, I find it to be so filled with human mediocrity that I wish it would stop. The doctor analogy is not a good one. I worked in a hospital for a number of years, and for doctors, you were either high caliber or you were let go. There was no room for sub-par people at a hospital. In the case of pistachios, that's another story. So, your choices are "great" pistachios at $14 a bag or good pistachios for $7 a bag.

Audiences want to see performers on stage and real actors in movies and on TV. They would see examples of what real humans could achieve, as opposed to humanoid artificial nobodies.
 
Last edited:
I worked in a hospital for a number of years, and for doctors, you were either high caliber or you were let go. There was no room for sub-par people at a hospital.

If that was *ever* true, it's certainly not true now. Hospitals are *filled* with sub-par nurses and vastly overburdened doctors. This is true in both the US and UK, where in many places you can spend a dozen hours bleeding in the ER waiting to be seen by people who barely speak the language. A friend of mine has been pulling her hair out dealing with medical professionals ranging from "awesome" to "why haven't they been deported" since October, for a problem that should have been resolved fairly quickly by competent medical care.


Audiences want to see performers on stage and real actors in movies and on TV. They would see examples of what real humans could achieve, as opposed to humanoid artificial nobodies.

ERRRRRRR.

View: https://www.youtube.com/watch?v=oyXl3IhonRM
 
And taking the medical bit out of the realm of personal anecdote, allow me to introduce you to Martin Luther King Jr/Drew Medical Center in Los Angeles. For *decades* it provided not only sub-par care... it was *allowed* to do so for political reasons. Many of the poor saps who got sent there would have been better off with an array of mediocre AI. "Mediocre" would have been better than "criminally negligent."

View: https://twitter.com/RichardHanania/status/1679970523013615618



A single "meh" Med-AI might not be so great. But you could have multiples. And many studies have shown the wisdom of crowds, when the crowds actually try. A dozen mediocre Med-AIs studying you could come up with conflicting explanations, but get enough of them, and if they can commune on the subject, they'll probably end up being better than a small number of experts.
 
Pfffffttttttt...
Is that the sound of air being let out of those who said that steam trains/cartridge arms/horseless carriages/photographs/smokeless powder/aeroplanes/moving pictures/semiautomatics/talkies/radio/telephones/television/internet/space tourism/LucyLiubots/desktop fusion/etc. are just fads that will never really catch on?
 
And taking the medical bit out of the realm of personal anecdote, allow me to introduce you to Martin Luther King Jr/Drew Medical Center in Los Angeles. For *decades* it provided not only sub-par care... it was *allowed* to do so for political reasons. Many of the poor saps who got sent there would have been better off with an array of mediocre AI. "Mediocre" would have been better than "criminally negligent."

View: https://twitter.com/RichardHanania/status/1679970523013615618

A single "meh" Med-AI might not be so great. But you could have multiples. And many studies have shown the wisdom of crowds, when the crowds actually try. A dozen mediocre Med-AIs studying you could come up with conflicting explanations, but get enough of them, and if they can commune on the subject, they'll probably end up being better than a small number of experts.
Injecting my personal experience as a resident alien green card holder living in Southern California near Coastal Orange County, the medical care available here is nothing but *spectacular*. I've had a (permanent, even though the doctor told me it could be removed, but the less invasive surgeries, the better, right? Feels like the real thing!) Titanium plate implant as a stabilizer for a broken wrist surgery, as well as a complete Titanium shoulder replacement (feels like the real thing again!). I've gone to our local ER for things as trivial as an ingrown toenail as well as constipation, and I've always had the impression that the medical staff were almost happy to have to deal with some utterly insignificant issue for sheer diversion, since the most urgent issue I remember being raised by the surrounding staff during one of my visitations was who would do a coffee run and what the individual order would be. If that "ER" TV show would have been based on our local ER, the show would have been cancelled half way through the first episode. The point is that the quality of US health care is *extremely* localized, rather than national or universal, because I shudder to envisage what my experience would have been with the same kind of injuries in say Kentucky.
 
Last edited:
I have to say that I am quite intrigued with this maniacal insistance on generating stupid videos when AI could do something more useful like cleaning the mess encoded in every botched suite of codes left behing by the software industry in the last decade, like fixing spelling correctors or with virus protection.

Why not a functioning Windows AI, a definitive Word, a sympathetic Excel help file, a backdoor free Firewall etc?

Flabbergasting.
 
The point is that the quality of US health care is *extremely* localized, rather than national or universal,
Yes.

because I shudder to envisage what my experience would have been with the same kind of injuries in say Kentucky.
You would very likely receive the same excellent standard of care... depending on *where* in Kentucky. The worst care you might get in Kentucky is also likely available in California, as the MLK Hospital story amply illustrates. The friend I mentioned currently dealing with the systems can tell you about the three or four hospitals she's been to in a twenty-mile area... one she has nothing but glowing recommendations, another she wouldn't send an enemy to. And as the Baltimore school system amply demonstrates, it's not simply a matter of how much money they get... it's who is in charge, how competent and corrupt they are... and what the local population demographics are, often enough. Regions that are heavily laden with, say, migrants of dubious legal standing often find their ERs loaded to overflowing with people who use said ER for all medical needs, because ERs *have* to deal with everybody, with the result that wait times can be tremendous, and staff become overworked and indifferent.

This is an area where Med-AI would be handy. A Med-AI would be region independent; it wouldn't give a damn about your financial status or your race, ethnicity, religion, political beliefs. It would just take the data and crunch it, and produce whatever results it produces.
 
I have to say that I am quite intrigued with this maniacal insistance on generating stupid videos

"Stupid videos" are the sorts of things that untrained humans can easily spot flaws with. the earliest AI vids were massively screwy; the latest ones are nearing flawlessness. The point of the exercise isn't the stupid video so much as the demonstration of progress.

Why not a functioning Windows AI, a definitive Word, a sympathetic Excel help file, a backdoor free Firewall etc?

Hop to it.

Also:

View: https://twitter.com/bennash/status/1758294394044649895
 
Last edited:
People are interpreting this wrong. In short: a peer reviewed scientific journal published an article with laughable Midjourney AI produced illustrations, including a mouse with junk bigger than the rest of the mouse. People are going on about how AI generated art will be the death of not just art but science publication. But here's the thing: the "peers" reviewed the article and allowed it through. it wasn't the art generation system that failed; it was the dumbasses who saw the art and said "yeah, sure."

View: https://twitter.com/GenelJumalon/status/1758189720528388458


*Anyone* should be able to look at this art and go "Hey, that don't seem right." But the "peers" didn't. This points out a problem lots of people have been pointing out in "peer review" for a long time.
 
Status
Not open for further replies.

Similar threads

Back
Top Bottom