I don't know what to think of this

Back from vacation and I discover this app MidJourney. A AI image generator. It generates images. from prompts. Like working with computers before windowing.
I asked for a caricature of an old man. It gave me these 4 options. In about 3 minutes. Note how beautiful the colors are. The one in the lower left is stunning. 

Is this an idea factory? Or should I just go fishing. I'm so intrigued on one hand yet horrified on the other. 
In the days before stock photos I made a decent living inventing chapters and scenes for comps. Often mimicking other artist or photographers styles. Basically toss away art. This is so imaginative that can hardly believe it. 

It disturbs me. How about you?



  • edited August 2022
    Seems to be where we're headed.  :/

    I can't deny that they are good. 
  • Kingston

    Spooky. I have to admire the quality of the characters pictured. But when you think that the graphics generator has access to lots of keyword tagged images. The result is probably a montage of typical old men.
    A bit too realistic for a caricature though. The images are neither comic or grotesque. The AI machine needs to polish up the concept of caricature.

  • Doesn’t bode well for graphic artists.  
    Have you seen the latest video that Mark did?  He seems to be getting into the metaverse.


  • I saw that, too, @GTO. Wasn't sure what to make of it. 
  • CBGCBG -
    edited August 2022

    AI is great at pattern matching and assembling multiple similar patterns from multiple sources.  

    It does not really operate conceptually although it works well with patterns etc that have been categorized.

    It cannot truly create  it can only recombine images it has processed.

    That is not to say that the human eye is not easily fooled by these images.

    There is no art added in the process of AI making these images.  A blind algorithm has no experience, cannot value, cannot find meaning, joy, sadness, in anything.  Any life in any image AI creates is in the original works which it recycled… and they are myriad and disjointed … they do form a cohesive whole.

    Art is individual and Human.

    IMHO this is no more disturbing than a 35mm camera.
  • If we are heading towards a world that considers there is no point in doing that painting you want to do because it can be generated in a few commands then that would be sad.

    Though I have a bit of a love hate relationship with regards to technology. I purposefully left my mobile phone behind when I went away on a ten day trip, bliss.

    Marks video was a surprise and just makes me a little more concerned regarding the future. When there will be no need to get off your arse and go and meet friends, go to a gallery or the cinema. Just pop those specs on and and you are there. In fact there will be no need for your to eat or go to the toilet as you will be on a commode and have feeding tubes inserted.  

    Its the future. Ha ha  :)
  • @GTO
    The meta_verse is not for me. I get sea sick with those headsets. The meta_verse is a different animal from AI though they may be in the genus.
    I worked on Madison Avenue when digital changed everything. I was lucky in that I had embraced digital 'desktop' production early. Later I worked in Financial news when the first digital events changed our lives and the lives of many on Wall Street. I was the guy who 'invented' digital charting for the WSJ. It changed the look of the old gray paper. I was told at every step that what we were doing could not be done. We proved 'them' wrong. I don't know that this AI stuff can't take over the conceptual image making 'market'. Even though the artist's doesn't lift a brush!

    I saw this today. Four or five descriptive prompts created this. Sure it's wonky but impressively wonky. Again lighting and color. Detail not so much.

  • @KingstonFineArt I’m sure the issue of lacking details will be resolved at some point. 

    @CBG I agree with you that AI cannot “create” art at this time but it does make me wonder a lot about the entire creative process, both technically and “sentiently”.

    @MichaelD I do find it odd that during this Anthropocene  period we are destroying, the ecology and some are looking for ways to escape via rockets to Mars and giggles to a digital virtual reality.  

    @tassieguy  I’m not sure either.  I don’t know if we can yet see the impact it might have both positive and negative.
  • edited August 2022
    Welcome to the Anthropocene. Wreck the climate, trash the natural world? That's ok.  We'll all live deliriously happy in a virtual world.  Well, I'm not sure that's a world I want to be part of.  Mark's video is about augmenting social media with VR - hang out with your friends at an art gallery, or any place you care to imagine. But where will that end? I suspect we'll end up with a large VR dependent population who can no longer function in the real physical world and who just can't wait to get home and get those VR glasses on again. Then they'll have  VR contact lenses so they never take the glasses off. It'll be the Matrix.  It's nuts. But it seems to be not too far away. Who will own and control the Matrix? Gives me the creeps.
  • I have a feeling that AI is one of those earthquake moments in how we work. Whether we are Dr., Lawyer, or artist. I've lived through 3 big shifts in art production. 1The shift from letterpress printing and hot metal type setting to offset and computer typesetting. 2 The shift from mechanical art production requiring many bodies to desktop art production. 3 The creation of very complex digital visual production on a single workstation. Namely internet cloud and video. Single point creation. 

    AI won't affect me too much except as a curiosity. I'm getting kinda old. But I can't help wonder how it will affect contemporary art. And if it gets smart enough even realism. Will it be a tool for inspiration when planning paintings. It's happening faster than I thought it would.

  • I imagine a future where the credit for creating art will go to the idea as opposed to the execution. 
  • Let me find an unexpected positive. @KingstonFineArt can you get it to generate forms of abstract art? Because throughout my lifetime people have argued that realism is not meaningful because we have a camera - therefore abstract in different forms are more valid.
    I don't accept that argument, but if AI can generate abstract art then those who argue this have shot themselves in the foot with that argument. Hoisted on their own petard.
  • When it comes to AI generated art.  I would expect artwork that uses an artists work as the basis for generating art without permission to be in violation of copyright laws.  

  • @KingstonFineArt That's actually quite pleasing. Well - there goes the 'no point in painting realism now we have cameras' argument. But there is a point in a person painting abstract even if AI can do this.
  • edited August 2022
    I don't think that copyright comes into play. The AI simply creates archetypes of styles and genres. Learning along the way. If you feed it an image it doesn't mimic the image.
    So far I see it only as an inspirational tool. If I were still in the editorial art biz it would be my right hand. This is just the beginning of this stuff. SkyNet.

    I just had Mid Journey make an image. skynet from terminator, surrealism, 4K, --ar 16:9

    click on the image and view in a separate window to see the detail.


  • jeez... the future is scary haha...
  • @KingstonFineArt copywriter law includes derivatives of a work or works.  If the AI uses an artist’s works to “train” the AI then creates art based on that data it has violated copyright laws.  If it has not come up in court yet I’m sure it will.  
  • Is this derivative?

    an oil painting of a gorgeous and hot aloha princess on a hawaiian in the style of alfons mucha, hawaii beach, tropical, flowers --ar 2:3 

    I could say it's derivative of N. C. Wyeth. Maybe any number of painters and illustrators from of the 20th century. The fact is it is 'influenced' but many images to the power of 10. The result is unique.

    I find these images impressive yet highly flawed. Too many oddities to use out of the digital box.
    I think it's a good thing though.  

    Look at the color in this image. The shadows in the face are magnificent. The garment! The impact. All sparked by a 23 word description!

  • GTO said:
    @KingstonFineArt copywriter law includes derivatives of a work or works.  If the AI uses an artist’s works to “train” the AI then creates art based on that data it has violated copyright laws.  If it has not come up in court yet I’m sure it will.  
    There is no art or artist without derivatives.
    I'm not saying that there is not some validity in this - but I've seen some complete nonsense court cases about 'stolen music', such as Men at Work's Land Down Under, demonstrating that copyright laws are sometimes very good for lawyers and greedy talentless people; and judges are sometimes completely out of their depth. So I agree. If it hasn't come up in court yet I'm sure it will. And if we tell you name of the game, boy, it's called riding the gravy train.
  • @KingstonFineArt her neck looks too narrow.  It’s almost like it took a cutout of her face and payed it on the narrow neck.

  • The images are all flawed. Too many dust specs. Long necks. Really bad hands. 
    That said the ideas are challenging. The abstracts I've made are road maps. I have been trapped but repetitive subject matter for some years now. I have no problem with execution but unique ideas. I mean how many landscapes and pears can you paint.

    The color understanding of MidJourney astounds me. Compositional cues are brilliant. But the images are highly flawed. The fresh concepts aren't. It's visual thinking. 

  • This is a representation of 'rust'.

  • @Suez,  as @KingstonFineArt, eluded to earlier, the change that we have witnessed in the past 50 years of our lives, particularly in the form of communication and AI, is authored by us as a collective (reference to the Borg - sometimes reality is stranger than fiction!

    e seem to evolve to that which makes things easier and more profitable for corporations, which we operate. If we are on the side of winning, getting positive recognition from our peers, creating wealth for our families, we feel that it's part of the necessary evolution and justify what we do. Win at all cost, we teach this in the best universities. Corporations look for the CEO that is capable and willing to 'make those tough decisions' to increase profits exponentially for their shareholders, which we are part of.

    Can we compare this conversation about AI generated art and fine art with fast food and home cooking? To cook a meal at home for a family takes time, money, knowledge, energy, and hopefully passion. And most families eat multiple times every day. It used to be that you had to cook to eat. 
    Top food corporations hired top chefs for R & D to develop 'fast food' products. The need to cook has been eliminated but of course there are many who still do for reasons of quality in many areas such as nutrition. There is also a cost factor, it cost significantly more to cook from scratch than eating 'fast food'. 

    The same holds true with art, I do not think that AI will be able to create quality art to the level of humans because it will never have human emotions. But that is not to say that the majority of people will care about the difference enough especially when you factor in cost.

    I wish art and artist's had a higher level of respect in society and more would be in positions of leadership. In general, the human qualities that everyone admires the most such as compassion and empathy seem to be greater among artists. Wouldn't we all be better off if decisions that had great societal effect be made from a perspective that included more of those distinctly human qualities. :)

  • I continue to be startled but the stylistic choices if the AI. The very limited palette of blues and ochres.

  • I think the ideas generated can be very inspiring. I tried to use the app myself but I can't get it to work via the Discord Server. I think the beta invite might have expired?

    After Photography, then Digital Art, I see AI as the next era in images.
  • I had a look at that, but no idea what I'm doing! :)
  • The model now is fee sharing. I think that’s unsustainable. There is an open source model where you would not have to share. It’s early times on this. 

    Digital copyright is a mess. Copyright in general is a mess. 

    The AI engines aren’t mature enough yet to think about it. They can’t even make hands. 

    Right now it’s inspiration only. 
  • Walked into a bookshop today and saw two new releases with midjourney art on the cover. This thing is still in beta right? Whatever fee structure they think they will have good luck.

    Weird times. Glad I didn't go into graphics program as a kid.
  • AI image from MidJourney software.

    A lesson from art school.

    We used Creative Illustration by Andrew Loomis as our textbook. In the abstraction process of creating an image we were supposed to make abstract light and dark shapes the might capture the reader. This image is nothing but abstractions cloaked as real. Created from this phrase 'criminal abank robber hides in an alley, smoking a ciggarette, hd, detailed,'.
    Not my creation or phrase

    Classical composition exercise. Amazing color choices. Really bad hand. No feet. Sooo much to work from.

    Comments on this image only please.

  • edited August 2022
    I'm not clear on who created the image. AI or artists.  However, for me it captures the "feel" of "criminal a bank robber hides in an alley, smoking a cigarette'. 

    Great colour and values. The smoking trash can (?) is amazing.  The abstraction works very well. The overall feeling is dark, sinister, seedy ... But it's a successful image, whoever created it.  I think it's great.
  • This is an AI. You can see the wonkiness of the back wall windows. The figure floats with no feet. The hand.!!!!. 
    The color is fantastic.

    Here's some more:
    flower from animal bones
    texture painted wall stucco adobe southwest colors detailed --v 3 --s 20000
    raining scene thunderstorm dark clouds dark colors environment distant mountains water color painting
    electron microscope mineral  crystal fractals
    dreamy white dancing petals of many flowers
    100 million bright horizon photorealistic highly detailed
    metal shield with a crab crest ornament hyper detailed dramatic lighting unreal engine 5
  • edited August 2022
    Yes, I saw the hand and lack of feet and the wonky windows but I still think the image is good. The errors don't seem to matter. The initial impact is exactly what was asked for. AI is getting very good. 
  • I just asked for this image;
    an artist palette, with mixes of red blue and yellow, Alponse Mucha, 4k, photo realistic, --ar 3:2
    It took about 4 minutes to get this...
    A wealth of color.
  • Amazing! Are we redundant already?
  • If AI hooked up to a 3D printer then texture could also be created like in real paintings.
  • @tassieguy
    We have been redundant for a long time.
    All the history and rules and methods of image making are available to us at any moment in time. But we don't have the bandwidth to access or assimilate it. AI does.
    If we can add to the resource maybe we'll become part of the product. I don't know.

  • edited August 2022

    So, @Suez, what, in your opinion, is the "overarching significance of such participation once the compartment doors are opened and the technological gold rush commences in tidal wave fashion"? 

    Or is this florid rhetoric just in service to another of those conspiracy theories certain types are wont to spout. 

    Thunder and lighting, Very, very frightening. 

  • edited August 2022
     @KingstonFineArt, I hope you are right that there might still be room for us, even though meat like us doesn't have much bandwidth compared to AI. That's an interesting take on the situation. But, in that case, I wonder who AI will produce art for. Us? Or itself?
  • @KingstonFineArt if you developed a painting from one of these images is that considered an original work?  Or ?

    just curious @KingstonFineArt what kind of image would get generated if you used the words ;  still life,  carved red bowl, Chinese boxes, terracotta solders, tiger striped background.

  • @GTO
    An interesting question. It doesn’t really give you exactly what you ask for. It interprets There are folks who are retouch and using images in all sorts of applications. There are programs that fix faces. It is early days 
    I do suppose if I got an image that I found to be mine I might paint it but I have dozens of images that are waiting for me already. And a fertile imagination. 
  • edited August 2022
    This is from a program call DALL-E using your request... still life,  carved red bowl, Chinese boxes, terracotta solders, tiger striped background. 
    These result took about 5 minutes.

    DALL-E Photo

  • edited August 2022
    Well, @KingstonFineArt, AI did what was asked of it and some of those images based on @GTO's prompts are not too bad. I guess with AI's increasing learning abilities it can only get better at making images we humans find intriguing.

    And I guess we can only hope that AI will be used to help solve our most pressing problems, such as feeding everyone and limiting the human caused climate catastrophe that is upon us, as well as for making interesting images. 
  • What would it produce if you asked it for a 'Rembrandt in the style of Vermeer'? :)
  • @Richard_P

    Im guessing a weathered craggy mans face with some nice warm light from a window to the left.   :)
  • @Richard_P
    The circularity didn't work but this did...A painting of Rembrandt by Vermeer

    The colors are astounding
  • Probably. It's AI. Not a school marm 
  • @KingstonFineArt thanks for doing that AI image generation. 🙏  I have to say though, I am not really impressed with the results though.  They look cartoonish and poorly arranged.  The AI needs to go to school and take some more art and design classes. 😀

  • Thank you! Interesting for sure.. :)
  • I see this stuff as sketches. Research. Thinking. Reference. 
    It’s up to me, you, to make art of it. 
    You, I, can’t take this stuff literally. 

Sign In or Register to comment.