I read several recent articles about OpenAI’s DALL-E 2, an artificial intelligence (AI) system that can create realistic images and art from a text description prompt. Example images are stunning. The machine learning system was “trained” on a data set of 12 billion images to interpret the natural language inputs and create corresponding images.
The system is not just a one-off curiosity. It’s the latest development in a long line of AI-created art. For example, a 2017 algorithm created portraits that were exhibited at the University of Edinburgh. And in 2019, an AI system called AICAN created an abstract painting that sold for $432,000 at a Christie’s auction.
DALL-E 2 is in a beta, invite-only state right now, so I haven’t been able to test it directly, but several other AI image generators are open to the public. Out of curiosity, I registered an account with MidJourney, an AI image system currently in a public Beta, and started playing around. I was stunned by the results.
With simple prompts, the system generated artwork of various subjects and styles. All of the images in this post were generated by the system.
I was amazed at how well the system was able to interpret my prompts and create corresponding images. And I wasn’t the only one. A number of other users have posted their own examples of AI-generated art, and the results are impressive.
So what does this all mean for artists?
Simply put, AI is coming for your job.
While the artwork generated still has a distinctly digital feel, the systems are in very early days and will only improve. It’s not hard to imagine a future where AI-generated artwork is indistinguishable from human-generated artwork. With already-existing printing technologies, it’s not difficult to envision a future where AI-generated art is mass-produced and sold in galleries and art fairs.
Of course, this is all speculative. It’s possible that the systems will never reach the level of quality needed to replace human artists. But even if they don’t, they will likely have a profound impact on the art world.
Consider, for example, the way AI is already being used in the music industry. AI systems are being used to generate new songs, and some artists are already incorporating AI-generated elements into their own work. It’s not hard to imagine a future where AI-generated artwork is used in a similar way.
It’s also not hard to imagine a future where AI-generated artwork is used to create fake masterpieces. With the right input, an AI system could generate a fake Rembrandt or van Gogh.
Before you retire your paintbrushes and give up in despair, I believe that traditional, human-generated art will always be sought after by art collectors. There’s something about a human touch that can never be replicated by a machine, and collectors will always be interested in meeting artists, creating relationships, and having a human experience.
In fact, despite the possibility of a flood of AI-generated art entering the low-end of the art market, AI may actually end up being a great tool for artists. Just as AI is being used to generate new music, it could be used to generate new ideas for artists to work with. And as the systems get better at understanding and interpreting natural language inputs, they could be used as a kind of artistic collaborator, providing feedback and suggestions to artists.
Here are five additional ways artists could potentially employ AI to help them create:
1. Brainstorming
Artists could use AI systems to generate ideas for new artworks. By providing a prompt, or even just a keyword, an AI system could generate a variety of images for an artist to work with.
2. Collaboration
Artists could use AI systems as a kind of collaborator, providing feedback and suggestions on works in progress.
3. Evaluation
Artists could use AI systems to evaluate their work, providing objective feedback on composition, color, etc.
4. Printing
Artists could use AI systems to generate high-quality prints of their work, making it easier to sell their work online or in galleries.
5. Marketing
Artists could use AI systems to generate marketing materials, such as website designs, social media posts, and email newsletters.
Are you nervous about AI entering the arts?
What are your thoughts on AI-generated art? Do you think AI-generated art will ever be able to replace human-created art? What do you think are the benefits of AI-generated art? Are you concerned that AI-generated art will take away from the uniqueness of human-created art? What do you think the future of AI-generated art holds? Share your thoughts in the comments below.
P.S. Over 60% of the text in this post was generated by AI
Visual artists aren’t the only ones facing the oncoming AI revolution. All of the green text in this post and the click-bait headline were generated by OpenAI’s GPT-3 AI writer. None of us is safe . . .
I can do a Mark Rothko in 15 seconds. Seriously.
If I could afford AI to do the characters in my novels, I’d save myself a lot of time,. effort, and ink outlines that have never been colorized.
It may look a Rothko from far away only, not close by
Stuff like Rothko isn’t quite a easy as it looks. And if you could afford to have AI do your novel’s characters, so would umpteen other people, people who would probably beat you to the punch that wouldn’t have even tried if they hadn’t had the AI.
I appreciate your take on AI and the approach that it can be another tool in the artist’s arsenal. Hear! Hear!
I think the whole process is fascinating, and if we embrace it and make it work FOR us, it doesn’t necessarily have to be against us artists.
I believe they said that photography would replace artists too. I remember when the first MAC computers came out and they said it would put graphic artists out of work. I spend a lot of time using computers to enhance my reference now and to catalogue my work. As long as we still have humans at the wheel, we can drive anywhere!
Photography is a different art form than painting–one that has it’s own demands in equipment and technique and can’t capture what isn’t there (like a dream or something that happened thousands of years ago). I got into photography just before digital took over. We used to have 2-3 professional portrait studies in town. Now, everyone just uses their cell phones. Photography is dividing into two camps–the people who can afford the fanciest “digital darkroom” and the worthless “everybody else.”
And I would point out that the combination of photography and airplanes DID effectively destroy the “bird’s eye view artists” industry.
I absolutely agree with Gina. The same doom and gloom was pronounced when photography came into being, the same when artists started painting on their tablets and when artists where able to make seamless collages on their computers. AI is just another tool. Sure, some will take advantage of it (think, the long line or artists’ forgeries over the centuries), some will embrace and explore it. I am confident that as we understand how better to control the technology (similar to learning how to play a musical instrument), we will be able to expand the tool beyond our current imaginations. I only feel excited by AI and where it will take me!
But depending on who you are as an artist, you will always feel like a cheat, and depending on how transformative you were with the AI’s prompts, and if you even added anything to it, or would have conceptualized it on your own – and therefore if you really had anything to do with , then calling it your own artwork.
Totally agree, ai prompters are cheats and aren’t artists… they’re just a client not the artist.
AI is coming whether we want it to or not. As AI depends on cliche — that’s how it operates, after all; it’s up to us to create unique and individual pieces.
I’m not sure how I feel about this. In some ways it makes me sad and a little freaked out. It’s already hard enough to be an artist in this world.
In another way, I don’t think that an AI will ever really be able to create the physical mixed media artworks that I produce. I already know that collectors are not interested in prints of my pieces, but want the originals I create with all the physical texture and layers that reproductions cannot capture. When I focus on that aspect, I’m not so worried.
Yeah, if you’re working int “2 1/2” dimensions, you’re somewhat protected from reproductions, but how many people are happy with purely 2D art that’s easily reproduced? The bottom line is AI art makes it possible for a lot fewer human artists to saturate a much bigger market.
I agree with you. I, too, employ a lot of texture and layering that would be very hard for AI, but, then again, they’ll probably be able to do that too. I’m counting on the value of human produced art to prevail.
AI art will never replace art created by humans. AI art lacks the one thing that makes us human: A Soul and a Spirit. I trust my own soul and spirit to guide me in what I create. It can change on a dime to head in a totally different direction that I could not even have imagined. AI does not have that spontaneity built into its program. It only follows what you tell it to do. It cannot rationalize what can come next and make changes in another direction when it is in the process of creating.
it’s humans that give art value, and a good PR campaign can inflate the value of anything. Remember several years ago, when an award-winning abstract was created by worms dipped in paint and allowed to wriggle over a canvas? Is that any more “human soul invested” than a guy who writes a painting algorithm?
Most of my work is in relief style or sculptural. I’m not too worried for myself.
I have a friend who is on the leading edge of AI work in Asia and it’s generating a lot of interest, but shares in tone what I see in the posted work here. It’s all very dense and heavy handed in how it uses colours. I haven’t seen any elegance or delicacy yet. When I do, l’ll have to rethink my view.
Have you seen what 3D printers can make?
Sorry for taking so long to reply. Just saw this. Yes, I have and some of it is extraordinary. Still haven’t seen an AI painting that grabs me though.
Although I’ve been aware of this trend for some time now, I mostly find it very sad. We are losing our opportunities to connect with other human beings in so many ways as it is. I think there are some things that AI should be kept out of, at least for the most part. Visual art, music, and writing would be included in those verboten areas if it was up to me. We are facing a crisis of loneliness these days, which I think will only get worse as the threats facing our very existence become harder and harder to ignore. Artistic expression by humans will be a hugely important part of our ability to get through tough emotional times, just as it always has been in the past. If we allow AI to take that over, we will lose something very precious about what it means to be human. I hope that doesn’t happen.
True collectors like originals. Too bad Al.
We’re already competing with China and artists from other regions where the cost of living (and taxation!) is a lot lower. For every buyer that’s willing to pay for an “original,” there’s at least ten that are perfectly happy with a print, or worse yet, don’t see anything wrong with photographing a work they like and printing it to a size of their choice. (Heck, I paint with a group that used to display their works at the local civic center–until they learned that one of the employees was taking the paintings to her home and photographing them for her own use because they “weren’t copyrighted”–i.e. as far as she was concerned, they were fair game because they weren’t registered by a big-name corporation.)
This is going to lead to more of the global elites owning everything and the 99.9% owning nothing, because the people who own these AI “artists” are going to be able to under-price the rest of the art world to the point where only the global elites can afford to pay a living wage for what will be a tiny pool of artists who have the connections to become celebrities.
Over the years of my life as I have observed the craft of “photography” evolve from what was originally a very manual endeavor ( from shooting to darkroom and everything in between ) requiring a number of skill sets, to the digital age where so many lines of what defines “photography” are now so blurred, I see the same type of trend now happening with these types of images. If I typed in “California coastal rustic barn in glowing golden field at sunset in the style of Edward Hopper watercolor” you’d probably come up with an image that’d look like a lot of my work. Perhaps I could use that image as reference for a watercolor. Perhaps someone would buy it off a FASO website I set up to sell just these kinds of images at any size on any surface. If I had the time and inclination to do that. This is just the technology that we have to use to our best advantage as visual artists, because it’s here, now.
Totally agree!
This reminds of a quote often attributed to Mark Twain: “The reports of my death are greatly exaggerated.”
Then there is another thing to keep in mind about making art. Art is either alive or it isn’t. Art cannot be conveniently pinned down or demystified, precisely because it is metaphysical; if its creation is magic, then the art is magic.
I also recall that art was given a death sentence when the camera was invented.
Agreed, Eden – it’s premature to call this the end of creativity.
As a graphic designer who is [surprisingly] still working in the field, this makes me sad. The MAC has taken away many jobs. I went to college in the 70’s and started working in the field when advertising art was a craft. Thumbnails, marker sketches, ordering typesetting, manual press-ready mechanicals, choosing special stock, press-oks… MANY of my contemporaries left the field then because they were just couldn’t or wouldn’t embrace technology.
Now, I basically do everything myself. And all of it at once. I build my electronic files from the get-go. I use my brain in a completely different way. I refuse to use “templates” because, I am being hired to give my client a “custom job” but that doesn’t mean they understand that the creative process actually still does take time and I can’t just sit at the computer and bang out something special for them in 2-seconds. My computer is a tool…. BUT…
With all the interest in “bespoke” goods these days: crafted IPO’s, farm-to-table, upcycled products… I’d like to hope there will still be a call for fully human-crafted, genuine art.
I think of it as a tool I could use. Photography did change visual art, but it certainly didn’t replace it. Using an AI interface will become another skill set. I agree it will be especially useful in the planning stages of a painting. I already draw digitally on different versions of my paintings to make decisions. I don’t think AI will replace human artists, but I do think it will be a useful tool. Also, it does add another layer of competition as it will certainly show up in a variety of venues. But I still see it as an amazing tool, nothing more. Someone still has to think up the prompts, and that in itself will become a new art form.
Zero Soul…..no depth ….
What a great tool for those “customers” who request a commission when they really want an artist to create what they see in their head. That type is a big headache who is never satisfied. Send them to an image generator and let them play all day!
If that Van Gogh rendition is the best AI can do, I am not the least bit worried.
This is the same kind of paper dragon as a Wal-Mart being built in your small city. They might be able to offer better prices at a lower cost, but humans still like the novelties and conveniences of small shops when shopping. The small business owner usually has little cause for worry.
The only potential problem I can see are crooked artists selling AI generated work to unsuspecting art buyers and passing it off as their own.
The quality will improve, Mason, but there will definitely be many ethical questions that will need to be answered along the way. What constitutes an original if it’s made in collaboration with AI?
I’d really appreciate having a “like” and a “love” to post for most of the comments here. As for my creating a look-alike Rothko, I’d find one of my other paintings that didn’t make the grade, and put 3 broad transparent stripes over it, allowing whatever it was to show through–thereby giving a failed work a compelling mystery. LOL
Pretty soon human beings won’t need to do anything, technology will do it all. With the soaring population growth throughout the world, I am wondering what purpose human beings will have other than creating more human beings. Human beings need to be needed, need to work, need to create, need to be around other people, need to have a reason to exist, need to feel proud of their accomplishments. So sad that things are moving too fast now. I wonder what type of world will exist for my grandson, who is 25, and for my children who are in their 50’s. Thanks for posting this news, even though sad and scary!
I’m an optimist Jaynanne – I like to think we are moving into a world of new possibilities. There might be some pain along the way, but there will also be new opportunities!
The text didn’t use the typical phrases that I often hear from you Jason. Intelligently written and clear, sure… but could’ve been written by anyone.
Today’s musical performers typically use auto-tune, even during concerts. I’ve come to recognize the sound. It takes vibrato out and audibly removes the human element. There’s very little effort for the singer to hit a note. Auto tune can make anyone sing on key.
When anyone can make art by words or sing on key with technology, it’s too easy. If everyone could easily create art, it loses value. It’s no longer scarce. Supply is everywhere, demand falls.
It could be a way for folks to make and print their own AI work, but I agree with you that collectors will want something made with human hands.
I love the beauty of a hand knit sweater. Machine made is too perfect.
Great perspective Lori – as everything becomes easier and more perfect, we long for the authentic. (This reply was written by a human 😉)
My brother has a painting of a grouping of flowers by an elephant. Looks pretty damn good and you talk about working for peanuts!
🤣
I don’t think AI generated art will ever replace human created art.
Mainly, that’s because in the past 30 years of my life, I’ve seen us all go through the transition from only mainframe, massive computers to the ones we now hold in our hands (smart phones and tablets), and I’ve seen ALSO the corresponding value placed on “face time” — the connection, in person, with a real human being.
Face time is the most highly valued commodity in every marketplace today, and purely physical objects, whether 2D prints or 3D printed, cannot replace it – even though I own several prints of classical works, I’d still pay my right arm and at least one eyetooth for the opportunity to correspond with or talk to Auguste Renoir, or Pablo Picasso. And I’ll still pay to go into museums that are exhibiting the original works, with pleasure.
I buy some terrific, high-quality recordings of performances of music and theatre to play on my high-end video and audio equipment — and still pay for tickets to live events of each where I can see and hear those artists doing their work.
We don’t need AI to print very high-quality reproductions of our art — we already have that technology in printers both 2D and 3D, and have had for 10 years now. No one has stopped buying originals, even if fewer people are doing it because fewer people have the disposable income to spend on them.
AI-generated art will always be different than what humans create, and humans will always crave interaction with other humans, both in day-to-day life and in their interactions with art.
So I’m not looking at AI as competition, not as remotely strong competition yet, and not as a replacement for human artwork ever.
I spent my adult career years working in the technology industry, and tech will never be able to completely replicate human-to-human interactions, no matter how “good” it gets at communicating with us.
SO I don’t accept that AI is any kind of serious threat to my later life career as an artist. And the rest of you shouldn’t either.
Excellent points Leah!
Reference those who mentioned or worry about the loss of human contact through technology.
Recommend reading, “The Naked Sun” by Isaac Asimov. It depicts a sculptor who creates in color simply by moving her hands, the AI does the rest. The story addresses the isolation of having AI doing everything for humans.
However, if I could sculpt the way described in that story….! You betcha!
I read the book many, many years ago -might be time for a re-read! Yes, looking at these as tools to be employed to further expressions seems like the most productive way forward.
Some of things come to mind.
AI is “Artificial Intelligence”. Where is the Artificial and where is the Intelligence? It seems this is worth exploring from the perspective of brain research.
In the article, you mention that the “program” is pre-loaded with images and words. That becomes the world the program inhabits. It’s a closed system. The only openness is putting elements together to create a third element from within the world.
Artists work from outside the “world”.
Not so fast you say. Isn’t that what we artists do? Yes but not quite.
The test would be to input one word or none at all.
What would happen in the artist’s studio?
What would happen in the AI program?
We humans, with our brains that are apparently not hard-wired, are constantly taking in new sensory information, creating an “image” of that information, evaluating how it fits with what we already “know”. The latest information on how memory is created is, it’s fluid not static. Over time, memory changes and our recollections become edited. Not only that, the remembrance of a visual scene may and usually does have an additional extra-visual sensory input with it. It’s usually not consciously attached but the thinking is, this additional sensory input helps to differentiate it from all the others.
All of this, can be machined so to speak.
What do you do for the visceral trigger?
It’s that prickly feeling in the pit of your stomach or the nape of your neck that serves to stop you and focus you. We all have a hundred ideas whizzing past.
What causes one to be noticed?
Why do we consider one idea over another?
Why do we expend time, energy, skill, materials, and emotion expressing our idea in material form?
What we artists know that all the AI people don’t, is art is an idea.
What we see as a result (the product of our work) is the clearest and best material expression of the idea. Built in, is the visceral trigger in the viewer that makes them stop and take notice.
We do what we do as artists because 20,000 or so years ago, an individual with an obsessive inner picture had to invent painting/drawing or go insane. We, today, can “See the horse.”
I use algorithmic programs for much of my work as an artist but it’s a tool.
I edit and adjust, erase, add a layer all on the basis of the communication loop between my inner idea and what I see.
I don’t think I’m much different form other artists in that regard. Finally, artists end up with a single image usually.
We don’t produce stand alone variations as a back up.
There are many questions to be asked about what the art process is by people who have marginally come close to what art is and what artists do that no one else does.
I view it as a tool. My buyers have expressed their opinion that they value original hand done art over anything else. I use wombo to work out ideas, and for amusement. I haven’t entered the NFT market or anything and don’t plan to.
I do feel it is a worry in circumstances as always with new technology and there could be some kind of cycle containing turmoil. The full circle being a return to the real value of the art and artist(signature).
What I mean is, when items such as art can be done at a whim and on scale, I think the soul would be diminished, materialised somewhat. Then again, that could be said about capitalist terms, when there is a frenzy and overpriced art (maybe I say this because I can’t afford it 🙂 ) and therefore given a value.
The biggest worry perhaps would be for the lower scale of art and artists, because our art will mostly be compared with the AI art by those who are not as educated or fascinated by artistic integrity, but by imagery. Not so much for the affluent collectors, wanting uniqueness, status and desirability.
So, those who would likely buy our art on the lower end of the bar may compare the both and go for the image. I have experienced this, having my original compared to the price of posters etc….Obviously, they must have been the wrong prospect.
At the same time, I believe there will always be those who appreciate and value artists much more, still, the threat will be present and could be damaging. One reason I say this is from a financial view, when income is crucial especially at the start of the career. For those that are established, they will be able to demand better prices and be in demand even. So I feel AI will hinder the ‘small’ artist until one gets a foothold.
Right now we are using technology to get our art across to people; online galleries, connecting with virtual and real galleries, POD and so on, but that is art mostly done by us, real artists.
Can you imagine when everyone gets on the band wagon and commercialises it all even further, more so by those who are not artists?
It is a big challenge, as always.
Nice to hear it could work as another tool for artists.
I
On the other hand I did read that Sophia the humanoid robot has sold a work for tens of thousands. I sell works too, but for peanuts still, in comparison!
So so elephants, so do chimps.
Maybe I should put on a gorilla suit too and see if that sells my stuff at the local art bazaar. Or dress up as an android.
Several comments have focused on the idea that collectors will always value the singular creation from the hand of a real human being. That certainly fueled the Arts & Crafts movement when factories began to turn out goods efficiently and at generally high quality.
The question I have is who are the collectors of today that will value the singular work in the 21st century and beyond?
Collectors of today have had a very different cultural connection to art than have the up and coming generations, who have been weaned with a cellphone in their hands and will text their companions while sitting next to them rather than communicate directly! Fast moving, virtual experience is increasingly trumping the real.
Already, interest in the art and collectibles of the past are being rejected as the historic ideas that animated them are being vilified by the educational/social systems in place. While bragging rights to ownership of a unique creation might seem relevant to every generation, the wealthy globalists of today are working to enable the “great reset” in which we will own nothing and like it–or so they say.
It seems that artists & gallerists will need to somehow make what we do relevant to these newer generations before we can count on them becoming collectors of “antiquated” art forms. We need to instill a love for these things as the finest expression of worthy culture that art has heretofore been.
I’m a professional art installer and already see that even very wealthy people use cheap art to decorate certain rooms.
Often there are one or two originals (for the living room and main bedroom) and the rest are prints or Hobby Lobby stuff. Not to mention non-art “art” like found objects. So the end isn’t near, it’s here!
But the artists who understand why people buy, and cater to this, will always survive.
These ai image generators are made with data centers that consume massive amounts of electricity, so much that its displacing people: https://www.ft.com/content/519f701f-6a05-4cf4-bc46-22cf10c7c2c0 . They work by scanning billions of images and blending them together; it is art theft. People are finding shutterstock watermarks and artists’ signatures in the outputs and this is just blatantly ignored because people believe the propaganda on how these things work.
This June there was a heatwave in India so bad that birds where falling out of the sky and animals where dying on the street: https://eand.co/the-age-of-extinction-is-here-some-of-us-just-dont-know-it-yet-7001f5e0c79a?fbclid=IwAR0xgSmvW9QO6AAeqgzjavoOJ6jBAAF_OCo2kQ5GlstEgF7csiLXyqV8jh4
In countries like Vietnam cropland is being poisoned by salt: as ocean levels rise the water moves farther inland and increases the salinity of soil, leading to dying crops and health problems for the people who eat the food that does grow. Just some of the current signs of climate collapse that people in developed countries completely ignore because its in the 3rd world.
You’re glibly speculating on a future that doesn’t even appear physically possible due to the many feedback loops of climate change compounding on one another and accelerating us towards total collapse. The technology itself, aside from being a gigantic scam that steals labor from artists in an attempt to replace them and transfer wealth to big tech, is also contributing to the total collapse of the environment, creating a future in which civilization as we know it won’t even exist in a complex enough form for automation to be even possible. You should view all this with nothing other than absolute hostility instead of this air headed consent-granting.
I’ve observed many technological changes in my life; born before PC’s, the internet and cellphones. Art, in my opinion, is communication, which humans seem to have had a need for at a basic level since before recorded history. My guess is, in spite of the fact that a group of eggheads designed image mashup software to regurgitate a print, it won’t stop artists from doing what they do, even if it turns out there’s little market for their work. 17,000 years ago the people painting in the Lascaux caves probably weren’t getting paid for it.
Possibly AI generated art will be a novelty for collectors at first once quality is established, but in the long run, I think human created art will survive if we as a species survive the mass famine that is going to take place in this century due the the catastrophic effects of climate change.
I am interested in testing AI to create several book illustrations. A book I am completing will require so many illustrations that producing them myself would take me a very long time, and might be impossible and paying illustrators might cost a fortune (that I don’t have). As both a writer and visual artist, I wonder if AI can produce decent illustrations from my own word concepts. And then I wonder can I manipulate the AI generations to tweak them to become better? We shall see!
I recommend giving MidJourney a try. It’s a bit complicated to get up and running, but it should start to give you an idea of what is possible.
AI – certainly an interesting technology that I see accommodating itself to say book illustrations – but to fine art? Don’t think so. (I’m assuming that these AI pieces are digitally generated then projected onto larger formats like canvas using oil/acrylic/charcoal etc.) Now let’s project for a moment into the future. It’s the 22nd century and we are in an art gallery listening to the curator talking to a client about the art piece he brought in to be evaluated. The curator is saying: Well Mr Smith I’m sorry to tell you that the painting was AI generated and not worth near as much had it been created with paint and brushes by an RP (real person).