• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Merged Artificial Intelligence

Roboramma said:
Why do you think that the highlighted will be true? Do you think it will be true indefinitely?
Click to expand...

Yeah I don't buy that.

In my favourite free art generator (https://pixart-alpha.github.io/), I regularly put in random words, phrases, nonsense words, or whatever song happens to be playing as a simple prompt, and most of the results straight off the bat have been really good if not fantastic, particularly the nonsense words, genuinely unique.
 
arthwollipot said:
They spent a fair bit of time talking about generative AI on this week's episode of The Skeptic's Guide to the Universe (live from Dallas) and Cara brought up something that I thought was a very good point.

AI will not take over from human artists. Humans will always create art - it's fundamental to human nature for us to do so. What AI will do is flood the market with mediocre art that is fit for certain purposes but in no way innovative or inspiring.

Further, I think that The Algorithm - itself an AI - will sort this out so it's not really a problem. Sure, YouTube will be flooded with crappy videos, but once people start clicking off them they will fall down the recommendations and disappear into ignominious obscurity. The good art - meaning the art that humans respond positively to - will be filtered to the top.

However, and this only just occurred to me as I was composing this post, AI that gets supergood at farming engagement will probably **** all that up. So maybe I'm wrong.
Click to expand...

They look at it in terms that are too extreme. As it currently stands, AI would simply do a large part of the work, while a human would do any necessary adjustments or even wholesale changes that their creativity demands. It's the people who like doing the AIs part of the creative processes, or are very good at that part of the process that would get shafted. While people who are bad at those things and better at others would benefit. Maybe someone has a great eye for perspective or shadows and light, and now AIs mean that this skill doesn't matter as much, or even at all.
 
Last edited:
Olmstead said:
Convince enough people (and perhaps the right people) that it's your artwork, and it's your artwork. Convince enough people that it doesn't matter, and it won't matter. Sometimes they might even be convinced for a good reason.

That answer is kind of depressing in its own way, but I think it actually applies to art in general.

Maybe no one else has thought of arranging teddy bears in this manner. Maybe using off the shelf teddy bears in conjunction with the AI tools of a major corporation is in itself an artistic expression of ... I don't know, the death of communism?

Maybe I'll print screen this image, watermark it with a bloody font saying THE DEATH OF COMMUNISM, and sell it for millions. It's not unintelligible, it's ... abstract.
Click to expand...

I've no doubt that I own the copyright to the line-up I did. That's clear copyright law, no grounds for it to be challenged. My point was more that I used assets provided by someone else as part of my artwork (legally I have a licence to use those assets in my artwork), when using an AI tool am I doing less as an artist so much that it doesn't count as artwork? What if I've used Adobe's algorithmic "spot heal tool" or their AI "remove tool"?
 
Darat said:
I've no doubt that I own the copyright to the line-up I did. That's clear copyright law, no grounds for it to be challenged. My point was more that I used assets provided by someone else as part of my artwork (legally I have a licence to use those assets in my artwork), when using an AI tool am I doing less as an artist so much that it doesn't count as artwork? What if I've used Adobe's algorithmic "spot heal tool" or their AI "remove tool"?
Click to expand...

I understood what you're saying, and I'm saying that whether it's an artwork is completely subjective, similar to the upturned urinal and other dadaist endeavours. Not saying this work is dadaist, but the same considerations apply both to it and other art. I'd say it's always art as long as even one person considers it art. Whether it becomes successful art would depend on who is convinced. The same applies to its commercial value. Although, who gets the money will be decided by courts, not necessarily based on artistic merits. Since you have the copyright, this isn't a problem based on current law.

We're discussing a throwaway forum joke, so it's a bit difficult to make useful statements about it. If I were an art critic in a hypothetical AI art world, I might point out that pretty shadows and other touch-ups are absolute garbage in a world where you can get that with a click of a mouse, so this isn't art. As far as the first image is concerned, there's been successful dada art using a similar cut-and-paste approach. You might be about a century too late though.

But let's assume you're not trying to get into a gallery exhibition, but just trying to make a buck. Well, the customer is king. I'M not paying for either of the teddy bear images, but maybe someone else would. I might buy some other heavily AI based art, but to my knowledge, I haven't as yet. Rightly or wrongly, I will also go out of my way not to, but that is just my choice based on subjective feelings. It's up to the AI artist to convince me (or perhaps trick me) into paying for their art, probably by making it really good or different in some way and beating the flood of other AI artists.

In the end, how much work one puts in, or how much art from other people one uses, or how much AI was used, really doesn't matter. Copyright matters for selling it, and that's completely removed from art and often decided by people who don't even care about art.

I have my own opinions about the copyright question and in which direction it should go, but to be honest, I don't know enough about AI to make an informed decision about it. Of course, the people who DO know enough about AI usually have a vested interest in making it succeed, so it's a bit of an unfortunate situation.
 
Roboramma said:
Why do you think that the highlighted will be true? Do you think it will be true indefinitely?
Click to expand...
Red and green are bright, vibrant colours. But if you mix them together you get a muddy brown, which is great if you want to paint mud, but it's not exactly an inspiring colour.

Generative AI flattens everything out. It takes human generated content and mushes it together to produce something that is better than the worst, but not as good as the best.

DeviantArt, once a vibrant artistic community, has become full of AI generated content, and quite honestly it all looks basically the same. Songs generated by AI are flat and sound mass-produced, like Stock Aitken Waterman at their blandest.

But as I said, humans will always produce art. It's one of the things that humans do. And no, I don't think AI will ever be able to produce works of the quality that humans can at their best.
 
That would require robotics, in my opinion, and in general people are a little shy of hooking AI up to robots.

Edit: can you imagine this with an AI controlling it?

 
Last edited:
arthwollipot said:
Red and green are bright, vibrant colours. But if you mix them together you get a muddy brown, which is great if you want to paint mud, but it's not exactly an inspiring colour.

Generative AI flattens everything out. It takes human generated content and mushes it together to produce something that is better than the worst, but not as good as the best.

DeviantArt, once a vibrant artistic community, has become full of AI generated content, and quite honestly it all looks basically the same. Songs generated by AI are flat and sound mass-produced, like Stock Aitken Waterman at their blandest.

But as I said, humans will always produce art. It's one of the things that humans do. And no, I don't think AI will ever be able to produce works of the quality that humans can at their best.
Click to expand...

That's the same as it has always been, good curating has always been neccessary. We should train an AI to curate AI art.... That's only slightly a joke - I think it will happen. Midjourney has a gallery and in that are some superb pieces of artwork. If you look away from the "hyper realistic" fantasy style that for some reason gets the most attention, you can find amazing examples that rival anything from a human artist. https://midlibrary.io/midjourney-style-classifier#styles-by-categories
 
Darat said:
That's the same as it has always been, good curating has always been neccessary. We should train an AI to curate AI art.... That's only slightly a joke - I think it will happen. Midjourney has a gallery and in that are some superb pieces of artwork. If you look away from the "hyper realistic" fantasy style that for some reason gets the most attention, you can find amazing examples that rival anything from a human artist. https://midlibrary.io/midjourney-style-classifier#styles-by-categories
Click to expand...

It's impressive, but I'd say it's not quite there yet. There are often still telltale signs, bits of anatomy that don't work or little details that don't make sense. For example, there was a chess board where the grid was at an angle to the boundaries of the board. And it wasn't just a checkerboard pattern, because there were chess pieces on it. AI recognizes what things tend to look like, but it makes mistakes like that because it doesn't know what it means.
 
Ziggurat said:
It's impressive, but I'd say it's not quite there yet. There are often still telltale signs, bits of anatomy that don't work or little details that don't make sense. For example, there was a chess board where the grid was at an angle to the boundaries of the board. And it wasn't just a checkerboard pattern, because there were chess pieces on it. AI recognizes what things tend to look like, but it makes mistakes like that because it doesn't know what it means.
Click to expand...

But you can say all of that about real art by real artists. In the showing off thread in forum community you can see some of the artwork I've done and that is full of little mistakes. I do agree that there are some cases where AI is still struggling to be convincing but that is getting smaller and smaller.
 
This is an area I've said before I expected AI to soon exceed humans- medical diagnosis, and it is happening now: https://www.cam.ac.uk/research/news...-doctors-in-accurately-assessing-eye-problems

A study has found that the AI model GPT-4 significantly exceeds the ability of non-specialist doctors to assess eye problems and provide advice.

The clinical knowledge and reasoning skills of GPT-4 are approaching the level of specialist eye doctors, a study led by the University of Cambridge has found.

...snip...

He added: “The models could follow clear algorithms already in use, and we’ve found that GPT-4 is as good as expert clinicians at processing eye symptoms and signs to answer more complicated questions.
Click to expand...

I'm not surprised about this; our medical professionals are not as good as we all tend to assume, especially not the likes of GPs, there is too much subjectivity in their day-to-day job.
 
AlphaFold 3 is now available, and it seems like some major advances since AlphaFold2:

https://blog.google/technology/ai/google-deepmind-isomorphic-alphafold-3-ai-model/#life-molecules
Our new model builds on the foundations of AlphaFold 2, which in 2020 made a fundamental breakthrough in protein structure prediction. So far, millions of researchers globally have used AlphaFold 2 to make discoveries in areas including malaria vaccines, cancer treatments and enzyme design. AlphaFold has been cited more than 20,000 times and its scientific impact recognized through many prizes, most recently the Breakthrough Prize in Life Sciences. AlphaFold 3 takes us beyond proteins to a broad spectrum of biomolecules. This leap could unlock more transformative science, from developing biorenewable materials and more resilient crops, to accelerating drug design and genomics research.
Click to expand...
 
It's a good one.

https://www.rottentomatoes.com/m/her

A sensitive and soulful man earns a living by writing personal letters for other people. Left heartbroken after his marriage ends, Theodore (Joaquin Phoenix) becomes fascinated with a new operating system which reportedly develops into an intuitive and unique entity in its own right. He starts the program and meets "Samantha" (Scarlett Johansson), whose bright voice reveals a sensitive, playful personality. Though "friends" initially, the relationship soon deepens into love.
Click to expand...

Although, having watched the YouTube video, I think that GPT-4o is not Samantha. It gets a bit cringey.

ETA: if you have Amazon Prime, I think Her is available there.
 
Last edited:
Puppycow said:


I've seen this movie. Have you?
Click to expand...

:)

That was one of ways I saw the technology going with assistants being able to hold a realtime conversation.

Still seems like not enough females in the developers, why have the female voice "flirt"with the male actor, why not the other way around, or the bigger issue, why flirting at all? Is that the only way they can imagine a "female" assistant being pleasant and causal with a male?
 
You must log in or register to reply here.

Back
Top Bottom