Is AI Art art?

When you are on the standard plan, you can switch to relaxed mode and keep creating without using fast hours. You have unlimited relaxed hours, it just takes a bit longer for the job to start.
Yeah I know, I never thought I'd use up all my hours tbh. I can't justify paying for the standard plan.
 
It is a shame these AI companies can't come up with a way of allowing just normal? people who'd like to play/fiddle/experiment a bit have access.
It's a bit like how I used to feel about Photoshop.

If I wanted to use Photoshop to edit a few family photo's, or if I wanted to run business, it's the same price.
I suppose like a subscription to Netflix, which I watch about 30 to 45 mins a day maximum, but I have to pay the same price as someone who has Netflix playing 24/7 on 3 TV's around the home.

wish there was a way of making this better
 
It is a shame these AI companies can't come up with a way of allowing just normal? people who'd like to play/fiddle/experiment a bit have access.
It's a bit like how I used to feel about Photoshop.

If I wanted to use Photoshop to edit a few family photo's, or if I wanted to run business, it's the same price.
I suppose like a subscription to Netflix, which I watch about 30 to 45 mins a day maximum, but I have to pay the same price as someone who has Netflix playing 24/7 on 3 TV's around the home.

wish there was a way of making this better
They have a free trial where you can make 25 images for free. And I think their cheapest plan is something like $10 per month. Even if you go for the more expensive one, you can cancel anytime and just have access for 1 month for example.

I'm on the more expensive one, but I treat it as therapy. Art therapy. I am able to exercise my creativity that was blocked for most of my life. I'm also terrified, because it is just a question of time before my skills will become obsolete thanks to AI. The world is changing in front of my eyes and it doesn't slow down for anything. So many people will be left behind, it's scary.
 
^ Indeed, I've sort of mentioned in Passing Ai at my workplace, and chat GPT, if anyone has hear of it, and the lack of awareness and interest is staggering.
Not even a glimmer of.... Oh, what's that then, tell me more.
No, just a blank stare and they typical it's just some unimportant nerdy thing reaction.

They really don't have a clue what's happening and is going to happen.
 
The new version is insane. I think photography is in trouble
Whilst the quality of the images is great, the content is still very random. It ticks the boxes of the prompt information, but it gives what it thinks is correct. All image briefs are art directed to the nth degree. AI cant follow a brief anywhere near closely enough for that. Someone penny pinching who isnt fussed may find it useful, but if theres a specific brief, forget it. Ive already seen art directors try it, and fail, returning to a human to get the work done.
It is extraordinary how AI image generation suddenly came of age, but its still got a way to go. Who knows what it will evolve in to.
 
^ Indeed, I've sort of mentioned in Passing Ai at my workplace, and chat GPT, if anyone has hear of it, and the lack of awareness and interest is staggering.
Not even a glimmer of.... Oh, what's that then, tell me more.
No, just a blank stare and they typical it's just some unimportant nerdy thing reaction.

They really don't have a clue what's happening and is going to happen.
AI's progress in just short few years is staggering. I simply cannot imagine what you will be able to do with it in the next 5 or 10 years. It's mind blowing.... and it is giving me an existential crisis already. I've been feeling like this for the past couple of days :)
 
Last edited:
Whilst the quality of the images is great, the content is still very random. It ticks the boxes of the prompt information, but it gives what it thinks is correct. All image briefs are art directed to the nth degree. AI cant follow a brief anywhere near closely enough for that. Someone penny pinching who isnt fussed may find it useful, but if theres a specific brief, forget it. Ive already seen art directors try it, and fail, returning to a human to get the work done.
It is extraordinary how AI image generation suddenly came of age, but its still got a way to go. Who knows what it will evolve in to.
The progress I see is mind blowing. While it might still be difficult to achieve a very specific result, for vast majority it might already be good enough. They have just released a new version a few days ago and it can be very hard to spot difference between a real photo and AI. Have a look above. Is it real or not? It's just a question of time before most of photography genres will become obsolete IMO.
 
The new version is insane. I think photography is in trouble
As long as it doesn't involve anything to do with fingers...
grid_0 (1).png
 
It's much better with fingers than before. Still not perfect, but a few months ago it was almost impossible to get realistic looking images. And now, this:

6aE0lKs.png
It's not reliable though which is the problem. I spent ages getting it to put out something I wanted, and did the upscale only for it to put out complete mince where the hands were. There must be some sort of "understanding" of what it's trying to draw, so is it really that hard for it to stick to four fingers and one thumb (they could even charge extra for the Polydactyly removal feature)?

The other problem I've noticed cropping up is "impossible walls" when it comes to buildings. It will start out OK, but as detail is added it will sometimes create a wall which couldn't be connected to the point it is. Initially it looks fine, but then you'll notice that the perspective of a door or a window is slightly off, and it's because the wall is doing two things at once. I've had to top up twice this month because I'm having to evoke --q 2, as it makes your time cost twice as much, but drastically improves quality. Well apart from fingers in my experience.

By the way, could you tell me what you wrote to get that output, just so I can get some tips please.
 
It's not reliable though which is the problem. I spent ages getting it to put out something I wanted, and did the upscale only for it to put out complete mince where the hands were. There must be some sort of "understanding" of what it's trying to draw, so is it really that hard for it to stick to four fingers and one thumb (they could even charge extra for the Polydactyly removal feature)?

The other problem I've noticed cropping up is "impossible walls" when it comes to buildings. It will start out OK, but as detail is added it will sometimes create a wall which couldn't be connected to the point it is. Initially it looks fine, but then you'll notice that the perspective of a door or a window is slightly off, and it's because the wall is doing two things at once. I've had to top up twice this month because I'm having to evoke --q 2, as it makes your time cost twice as much, but drastically improves quality. Well apart from fingers in my experience.

By the way, could you tell me what you wrote to get that output, just so I can get some tips please.
I'm a cheapskate. I use my 15 hours of fast time but the rest I just use relaxed mode. I've been using it so much this month that they have slowed my relaxed time down artificially and I have to wait 5 min each time for the job to start, but I still refuse to buy more hours :D

I've not done much of architectural stuff TBH. I'm more interested in people photography, as this is where my job is and this is what will ultimately make me jobless.

I have joined MJ a few months back and at the time the realism wasn't there, but in just a few short months it went from sort-of portraits to ones that you really have to look closely to distinguish between real and not real. I cannot imagine what it will be able to do in another year or 5 years. I, for one, will be one of those people affected by it. I'm terrified, but I'm also fascinated and mesmerised. It's like watching a car crash live in slow motion, it's awful, but you cannot look away.

I think this prompt I had AI to create for me. It's an old one:
advertising photograph featuring a young woman in a confident and relaxed pose, wearing fashionable clothing and accessories that complement her natural beauty. The lighting and composition could be carefully crafted to create a visually striking and attention - grabbing image
 
I'm a cheapskate. I use my 15 hours of fast time but the rest I just use relaxed mode. I've been using it so much this month that they have slowed my relaxed time down artificially and I have to wait 5 min each time for the job to start, but I still refuse to buy more hours :D

I've not done much of architectural stuff TBH. I'm more interested in people photography, as this is where my job is and this is what will ultimately make me jobless.

I have joined MJ a few months back and at the time the realism wasn't there, but in just a few short months it went from sort-of portraits to ones that you really have to look closely to distinguish between real and not real. I cannot imagine what it will be able to do in another year or 5 years. I, for one, will be one of those people affected by it. I'm terrified, but I'm also fascinated and mesmerised. It's like watching a car crash live in slow motion, it's awful, but you cannot look away.

I think this prompt I had AI to create for me. It's an old one:
advertising photograph featuring a young woman in a confident and relaxed pose, wearing fashionable clothing and accessories that complement her natural beauty. The lighting and composition could be carefully crafted to create a visually striking and attention - grabbing image
I used to work in IT, so I'm more likely to pay for stuff when I see the work going into it, that's purely the reason I pay for it. I need to check the relaxed mode for paying users. I'm sure it comes out of the time at 50%, and if so then I'll be unsubscribing because nothing I generate is time critical.

As for photography dying, I don't think it will. If you look at things like stock images, you can usually tell what they are. They're just generic images that have been shot as part of a series, and they'll have been touched up or whatever. They're sterile because they're too perfect. A real photograph that people have taken time to get the perfect shot with, or are "in the moment" (like a F1 race) has that feeling. It's almost tangible. AI doesn't have that. It's making images, but it's training itself to be perfect which will result in stock images. That perfect shot is something that happens, and is not down to an algorithm. Yes, the image you've done is great, but I don't "feel" it if that makes sense.

It's like me with buildings and signs. It allows me to throw myself back to a time I wasn't in and imagine what it was like. I don't want conventional photorealistic, I want that faded sepia image that the photographer struggled to set up using a big tripod camera. I want that little amount of blurriness, and the natural noise and dirt on it (not generated by an algorithm). All the little flaws contribute to it, as clichéd as it may sound.

I can use ChatGPT to to write code for me, but that comes at the cost of it doing it properly. I can write some code that's slightly hacky, and not the proper way to do something, but there's not as much of a performance hit because I know what the output will be. Getting an AI to produce code where it can use hacks or workarounds is problematic unless the AI can fully understand and think "outside the box". A lot of it can be worked out in places like the shower or drifting off to sleep (not good, because then I want to test it), or by looking at it from the outside. Even video game developers will use hacks to make the system do things it strictly shouldn't (or at least without a hit).

The real question I suppose, is how do you explain what causes a "Eureka!" moment to an AI, let alone how do you code it? That initial idea may be wrong, but through testing it and refining it you have that breakthrough. AI will test it, see it's screwed up and sack it off as "it didn't work", while humans can go "OK, so this does that, and that does this, so...". We're able to analyse unexpected results, and use imagination to figure out what is wrong. Sometimes the undesirable results can lead to spin-offs in novel areas, and it's because humans can link unrelated things together. AI can't step outside it's programming, humans can.

Thanks for the /imagine prompt. It looks like I might have be a lot more verbose in regards to what I put.

One of my favourite own generated images recently has been Woody. It was just given the prompt nightmare fuel.

nighmare1.png
 
Last edited:
The progress I see is mind blowing. While it might still be difficult to achieve a very specific result, for vast majority it might already be good enough. They have just released a new version a few days ago and it can be very hard to spot difference between a real photo and AI. Have a look above. Is it real or not? It's just a question of time before most of photography genres will become obsolete IMO.
its undeniably a great image, but its definitely not photo real. It looks like a CG render, by a talented digital artist trying to do a photo real image, but not quite getting there. Probably because the AI has scraped places like Artstation, and not photo libraries. Or maybe it just creates detail that has a look similar to rendering software?
Nothing looks photographic, the way the skin folds and wrinkles looks like its the work of someone who is good at the technical aspects of computer graphics, but lacks the hard attained observational skills of an accomplished painter. Ditto the hair and generally unconvincing chiaroscuro.
Ironically, the strengths of AI image generators, is the way they can think outside the box, come up with ideas that are unexpected. Having said that, I can now often spot an AI image a mile off. This may be offputting for it to be used in a commercial setting.
the fact that its an evolving technology may mean it gets better in time.
 
Last edited:
The images in #56 and #59 look like pictures to me, admittedly with a filter for #59. But then I'm not a photographer or graphic artist. I would say most of the audience for these images would think them real too as really they wouldn't be looking too closely.
 
I work with commercial photography mostly, so that's what I've been testing. The image of the lady model is meant to be an advertising photography, hence it looks maybe overly perfect, but you can easily make normal looking portraits with the AI engine as well. I've been playing with it a bit more and I'm impressed. Of course, would that kind of images be enough? That's the decision of each individual client, but especially the ones with lower budget might be tempted to forgo going professional photography route.

It takes a team of creatives at least a couple of days to set up for a shoot like this. Planning, creating mood boards, deciding on wardrobe, makeup, hair, building set, setting up lights, test shots and then we can create. Then there is the post production stage. It can take even a week or more to get a single finished, polished image.

With this, all you need is a single person, keyboard and a subscription plan and in 10 minutes you can have dozens of images to choose from. Yes, it might still be a hit and miss at the moment, but for 50 images like that, they will be able to choose at least a few that are OK I think. And you can keep going for an hour or more and have hundreds of images to choose from in time where it takes just to do makeup and hair.

I'm not predicting that photography will be dead soon. There will always be photography as long as there are people who want their faces on images, but I'm predicting it will impact photography in a way. Some genres of it more than others, but I think it will have impact the commercial side at least.

Even though I say that, I still think this is a great tool. Even for creating mood board images. I don't have to scroll through endless images online to find something I like to add to moodboard. I can simply write what I want and have plenty of images to choose from.

Besides, you guys judge the images as they are now. I see it as just a snapshot of a moment in time. The images look like that today. How will they look in 5 years time. Will you be able to distinguish real from not real then?
 
Last edited:
For moodboards, yes they are excellent.
From my experience, an art director would reject everything Ive seen so far from AI.
To a casual viewer, it looks like photography. But under the scrutiny of an art director, they would point out all sorts of issues. Primarily, the overly perfect airbrush like detail. They look like heavily retouched photographs, that stray in to phone app filter territory, not up to commercial use.
Currently, its not at a point that the majority of clients would accept.
However, Ive no doubt it will evolve and improve, so who knows where it will lead.
 
For moodboards, yes they are excellent.
From my experience, an art director would reject everything Ive seen so far from AI.
To a casual viewer, it looks like photography. But under the scrutiny of an art director, they would point out all sorts of issues. Primarily, the overly perfect airbrush like detail. They look like heavily retouched photographs, that stray in to phone app filter territory, not up to commercial use.
Currently, its not at a point that the majority of clients would accept.
However, Ive no doubt it will evolve and improve, so who knows where it will lead.
I agree that it can be a bit heavy handed when it comes to skin, but let's not forget that in commercial photography skin is usually heavily processed. Yes, there is trend for more natural at the moment, but it is still far from really natural skin. Once I get some more free time I will experiment to see if I can force it to have skin more natural. But the difference in skin texture itself between a month ago and now is amazing.

As for an art director rejecting? I guess it depends on the art director and budget. If your budget is on lower side, you might not have the flexibility of a large production. Travel overseas to set up a shoot it expensive, especially if you only need a few pictures for your website and promo leaflets. No you can have them for $10 per month. However many you want and from every location you can imagine.

Just a quick try to get more natural skin



commercial stuff



commercial portraiture
 
Last edited:
I agree that it can be a bit heavy handed when it comes to skin, but let's not forget that in commercial photography skin is usually heavily processed. Yes, there is trend for more natural at the moment, but it is still far from really natural skin. Once I get some more free time I will experiment to see if I can force it to have skin more natural. But the difference in skin texture itself between a month ago and now is amazing.

As for an art director rejecting? I guess it depends on the art director and budget. If your budget is on lower side, you might not have the flexibility of a large production. Travel overseas to set up a shoot it expensive, especially if you only need a few pictures for your website and promo leaflets. No you can have them for $10 per month. However many you want and from every location you can imagine.

Just a quick try to get more natural skin

9FJ4W9l.png


commercial stuff

aeox4CB.png


commercial portraiture

mA8cR5B.png
these are all amazing images. All the more so considering how they are created. But they are still very much high end computer graphics render quality. Its amazing that this exists, ive no doubt theres a use, but for my line of work, they arent of a quality that is usable. It seems AI is still having difficulty in rendering truly photographic images. Ive no doubt it will come. I kind of think it may not be a technology issue, more a coding issue. The software is written to produce this kind of image quality. It may well be deliberate, as its very on trend currently to have super smooth detail in portrait photography. The Kardashian effect.
 
these are all amazing images. All the more so considering how they are created. But they are still very much high end computer graphics render quality. Its amazing that this exists, ive no doubt theres a use, but for my line of work, they arent of a quality that is usable. It seems AI is still having difficulty in rendering truly photographic images. Ive no doubt it will come. I kind of think it may not be a technology issue, more a coding issue. The software is written to produce this kind of image quality. It may well be deliberate, as its very on trend currently to have super smooth detail in portrait photography. The Kardashian effect.
I guess it all depends. There are so many sub-genres, even in the commercial field. Some like the skin to be as unobtrusive as possible, some more natural. It all depends. I bet in future, there will be more control on how you create AI images and results will be less random. Please note that I have not specified how the skin should look in the last two images. I only did in the first one. I need to experiment some more with it to see if I can get the images to look a bit more natural.

For now, I think if they can keep that speed of a progress going, it can change the way we create images in future.
 
Well, I had a go at my creation, and you may question the subject matter ;)

bea78ffd09f6b7ca93ed077586445a4a.jpg
 
The filing of the teeth to points is perhaps a rare insight into the designer's psyche...or not.
 
This illustrates what AI is good for.
Wierd eyelids, pointy teeth, useful ideas for some kind of alien, vampire, character design.
 
And this also was created by AI:

Cheese has been an important part of human cuisine and culture for thousands of years. It is a versatile and delicious food that can be enjoyed in many different ways, from a simple grilled cheese sandwich to a fancy charcuterie board. Cheese is also a rich source of nutrients such as protein, calcium, and vitamins, making it an important component of a healthy diet.

Beyond its nutritional value, cheese also plays a significant role in cultural traditions and celebrations around the world. Many countries have their own unique types of cheese that are an integral part of their national cuisine, such as cheddar in the UK, brie in France, and queso fresco in Mexico. Cheese is also a common ingredient in many popular dishes and cuisines, from pizza to pasta to burgers, and its flavor and texture can greatly enhance the taste of these dishes.

Overall, cheese is an important and beloved food that is valued not only for its taste and versatility but also for its cultural significance and nutritional benefits. Whether enjoyed on its own or as an ingredient in a recipe, cheese remains a staple of many people's diets and will likely continue to be a popular food for many years to come.
 

The latest video from AVForums

Is 4K Blu-ray Worth It?
Subscribe to our YouTube channel
Back
Top Bottom