Artificial Intelligence is making leaps and bounds on many fronts. Recently, AI Art hit the internet in a huge rushing onslaught of images as people climbed aboard and experimented with the tools. Midjourney was the first one I tried. I didn’t create anything spectacular and didn’t save those images as I was mostly just curious at the time. In those early days, I saw more architecture AI Art imagining new structures that were very curious and interesting.
As well, dresses that seemed to come from fantasy movies were another one that I saw a lot. Since then, it’s expanded dramatically and a wide variety of AI Art has hit social media.
December 20th, 2022 I shared Topher Welsh’s AI art created on Midjourney of re-imagined superheroes as babies. It was comical and really quite incredible. Our Facebook page Arts Artists Artwork leans toward more traditional art and hasn’t shared the AI Art extensively like other pages have but I made an exception with the superhero babies because I thought they were cute. I was also very curious how the audience would respond and considered writing about it. I didn’t expect the backlash the post made though, as numerous artists expressed a sense of betrayal and were surprised our art page would share anything created by artificial intelligence.
As of today, the date I’m writing this a couple days afterwards, I’ve discovered some of the more intense discussions were deleted by the originating commenter. That puzzles me, but I did secure some healthy conversations with some follow-up posts about the movement NO to AI ART
…and a quote meme about AI Plaigarism among AI creators.
In the course of the discussions on these posts, I solicited two active participants who had a lot to say. I asked them to write a contribution to this article in response to what is happening with AI Art and the issues and opportunities as they see them. First up is Maxim Lamothe who is both an artist and a computer scientist. And second is Matthew Lamm who is both an analog and digital artist who has been very wary about publishing his art publicly online for easy consumption. Before I introduce their contributions, let’s also look at Cat Fink’s comment about copyright and AI.
Cat Fink’s Comment about AI Art and Copyright
AI Art tools are here to stay. The true misfortune here is that many artists are unaware of how copyright laws work, and the consequences of posting content publicly on the internet. It doesn’t matter if something is copyrighted or not—machine learning is currently legal. The courts also recognize that not all aspects of an image are subject to copyright, which means bots can legally copy certain data. At this point, it seems like everyone is following the rules, although there will be a few cases that are tested in court.
Copyright practices established during the early days of the internet may have failed to keep up with advancements in machine learning. However, it’s important to note that the current “gaps” in copyright laws that allow for machine learning and art generation also protect all artists during the creative process. Copyright law recognizes that all art builds upon other art; so ideas, concepts, systems, or methods of doing something cannot be copyrighted. Without this freedom, it would be impossible to learn or create new works. This is not just a freedom, it is a right belonging to all of us.
I know many people want to ban AI, but it’s going to be a negotiated balance between those in favour and those not, because both parties have rights under the law. I support ethical use of AI; but ethics are the choice of the individual, and are not regulated by law. Therefore it remains up to artists to educate themselves on the limits of copyright law, and the risks taken when uploading images to the internet. Opt out scenarios and other proposals where artists have more control over their data need to be explored and pursued.
A.I. companies and end users are making voluntary ethical changes. This needs to be recognised by Anti-AI groups. Also, Anti-AI trolls need to stop harassing AI users and spreading disinformation. If one group of artists continues to exclude and harass artists who use A.I., then A.I. users will be less likely to adopt ethical practices. Ethics are VOLUNTARY, so we need unity and inclusion in the art community.
If you really want to understand how the courts see machine learning, and current thoughts on legal matters, I highly recommend reading this article from the Texas Law Review.
About Maxim Lamothe
Maxim Lamothe is a music enthusiast playing electric guitar and was hunting for contracts for an underground band before the associated band stop producing music and disbanded. He also enjoys abstract pencil drawing on paper and keeping them in a protected album. He also had the wonderful chance of learning computer science which got him to understand how Artificial Intelligence works on a surface level fairly quickly. His artist name is Ishio Ramoyald.
Maxim Lamothe’s Response: An Opportunity For Artists With AI Art
I was fortunate enough to bait an actual artist that had a head on his shoulder to expand a little bit on my take on how to apply AI art to help artists make more money and I’ll be adding it to my overall commentary here.
1. What is AI Art?
AI Art is art made from Artificial Intelligence. It uses something called a dataset to learn how to create artistic creations through coding. It’s a first for Digital Assets creation and many people have concerns about the ethics of its usage.
2. Why are people concern about AI Art?
The concerns of many artists come from how they are building a dataset. You need data and the data currently being used to create them comes from copyrighted materials. Since there’s no way to protect their artistic creations from Web-Crawlers, artists, photographers and other creators find themselves powerless toward this new technique for creating arts. Many of them invested their whole life to hone their craft and it’s being stolen from them.
3. Can we fix how dataset are created?
I do think there’s a possibility to create a private dataset specific for each artist and they could monetize those to the users but we first need to determine the legal aspect of selling a completely open-source software for commercial purpose. This is quite a headache as this new structure needs to be talked in length between the artists and the Developers (Devs) and, as of today, we haven’t seen any move on both sides.
4. How could we get artists and Devs to talk to each other?
Devs will listen to logic that won’t hinder their product from being effective, Artists want their protected work to be paid for. Now that some private paid solutions are becoming more and more available, artists should get a compensation if their arts are at the centre of the creative process. Many users of AI Art bots are explicitly using artists names to get specific results and this can be problematic in the long run.
5. How can Artists better protect their art from Web-Crawlers?
I will use my own case as an example to better illustrate what I mean by protect your art. I’m not an artist who produces to make money in the first place but I do understand how easy it was for humans to grab my creations and use it for themselves if I wasn’t careful enough about where I want to put it. We need to rethink the way art is being sold to others so that once the art is sold, it’s a done deal. It doesn’t belong to the artist anymore; it belongs to the one buying it. The current system is not compatible with how AI Art works and terms of services for both artists and buyers needs to be tweaked quite drastically.
- Solution #1: Don’t put your art on the web until it’s sold, keep it out of public eyes and use a unique sample to showcase your skills. Once the interested is picked, work on a commission or let them access a gallery where you control 100% who gets in and out and make sure you are watching what they are doing. Always assume people are there to steal from you.
- Solution #2: Devs could work on how the Web-Crawler gather data and when it finds a specific code embedded in a picture, it will NOT grab it for training it’s dataset. This would be a code made available by the Devs of the AI Art bots either under an Open-Source license or a Paid Licence depending on the popularity of the tool they are using. For example, Stable Diffusion would use a code that is Open-Source and MidJourney would sell their own version for their modified version. (since money is involved, I expect artists to complain about this solution)
- Solution #3: Paid Dataset. Each artists have the liberty to create their own Datasets and only people who pay for those datasets can be used. They would be infected with a self-destruct code that can be activated even when offline to try to curb the attempts at piracy. This would be a very complicated setup but it could be a possibility if the Devs and Hackers are willing to work together to make it happen. (Not all hackers are black hats)
6. Can Watermarks/Signatures product be a good solution?
Unfortunately, still as of today, OCR technology still has issues with characters on different coloured background and all the OCR technology is very expensive. This is unfortunately too costly to implement at a large scale in my opinion. I could be wrong and wanted to state it but so far I don’t see it as a valid solution. This is why I’m more opting for a code that is embedded in the pictures themselves instead, a custom-made solution from the Devs to the Artists that is fairly affordable.
7. Who should we select to talk about this issue?
We need as many people as possible from all range of expertise to offer their feedback but it must come with a risk to prevent trolling and false feedback to bleed into the conversation. We would need one moderator for each member on the discussion panel to prevent bad actors from derailing the conversation into asinine conversation and rules should be set in advance on how to exchange at the table like any good debate. Since artists are the most affected by this new technology, they should be the one controlling the arena while the Devs should set the rules. This comes back to the point two, Devs uses logic, Artists uses emotion kind of deal. If Artists are in an arena where they feel secure to talk, the Devs will be willing to exchange as long as everyone follows the rules.
8. Why invite anyone to the table?
There are Devs who really care about the Artists struggle and actually want to find a solution that will work for both and there are Artists who are already using AI Art for their own without seeing many issues. The more perspective on the subject we can bring to the conversation the better and we need to show that people on both sides want a solution. Having different levels of expertise will help the less knowledgeable users understand the troubles it brings to people with more popularity and fame without having it. Having hackers is a must in this situation because any solution that will be brought can be exploited and we need to figure out fixes for those before they get out of hand because of criminal actors.
9. Who should we select for the moderation/mediation?
We need a mix from both sides and they need to be extremely well versed in psychology to keep everyone on the subject. There’s no point in selecting someone that is not used to managing a big amount of toxicity skilfully otherwise they will be easily exploited by the members at the table to take a side. Impartiality on the subject is important for the selection. This is the part that will take the longest to do as you need to audition a lot of popular figures and some of them will simply refuse to contribute without money incentive.
10. Why are we talking about this?
The longer we take to start the discussion about the ethics of technology, the harder it gets for both sides to want to hear each other. Toxicity is already taking grasp in the minds of both sides because they feel either attacked or ignored. Public debates have been frowned upon for a long time as we could see with other technology advancement like Neurolink, mRNA, Cloning, chatGPT, digital ID, Cellular Towers, etc. There is real danger associated with unethical uses of technology and we all need, as a world population, to start taking the progress of technology way more seriously before it takes a big hold on all of our lives.
11. Where should we go to start the process?
Arts Artists Artwork has brought the subject to the public eye and is willing to help make your voices available to start the discussion on what solution could be viable, who we should contact and how to proceed forward with AI Art. Do not contact them for other technology subject mentioned in the previous point as they were just examples of other ethical subject that needs more attention. You can add your comment below on this article blog post. You can also join the conversation on the No AI Art Facebook page post. Trolls and bad actors won’t be tolerated and this rule will be very strict.
About Matthew Lamm
Matthew Lamm is both an analog and digital artist, also with extensive experience in photo retouching. He is an amateur art historian but endeavours to unearth art history’s lost artists & lesser-known works. He works at keeping up with countless emerging artists earning their place in history today and tomorrow and has a small but growing collection of living artists’ works.
The work Matthew Lamm produces is largely removed from his real name and operates under an anonymous pseudonym, much like our pal Banksy, and is very active. Separating personal from professional identity struck him as having more pros than cons, liberating the inner artist from the limitations of who other people think he is.
Matthew Lamm Responds to Address AI Art
Trying to push beyond my bias is a challenge, but AI’s best points:
Good for composition planning and idea generation. It’s very quick and responsive and usually gives the user a polished product from the start. But being too polished too early can also limit direction. If the goal is just to get some flashy imagery but often with a compromise to accuracy and funky glitches, AI will deliver on that. It can easily produce viable commercial products, provided it doesn’t hit a snag on rendering something it doesn’t understand, like fingers or toes. So one might retouch the images and correct for the defects manually, but are they then entitled to call the artwork their own, having altered just the small details by hand? Still very shaky ground, because what if the face is still 50% or more taking from another artist? The AI user would never have a clue, having no intimate awareness of the images used in the dataset. But a trained artist that’s spent a lot of time in the art world and actually seeing many of images firsthand just might, rendering anything produced by AI of questionable legality. What if the AI could actually break down for the user what percentage of images its dataset is using, to what degree? When we lay out some hard numbers, I think understanding the issue of legality risks becomes crystal clear.
I’ve seen AI kinda shake up the art world a little bit and produce pieces very unlike we conventionally see, some of which I’ve downloaded or can probably source. Some remarkable Art Nouveau architecture and interior design particularly caught my eye. I think AI inadvertently dropped hints about what makes art cool without doing it perfectly. How it might reinterpret famous historic artists and their work is always a nagging curiosity, but it’s an itch I won’t scratch after my trial experience because once I understood the patterns of AI, I could more or less predict its output. And with enough information gathered after testing, you can teach humans to get AI-like results without using AI, if that is the aim.
Still, how would AI interpret M.C. Escher? I’m sure I won’t have to wait long to find out, as there’s likely several thousand prompters asking that very question right now. But I know for my own account, I’d rather understand how I would interpret Escher organically. What if I wanted to merge Kadinsky and Escher and cultivate the mix as my abstract style? AI could probably model a good blend, but had it done all the work for me without me even trying to use my imagination, I’d be cheating myself out of copyright I actually could call my own, and seeing AI’s suggestion would just slay my own originality right away. It’s like having an art director over your shoulder throw you ideas you can’t use because the origins and application is ethically questionable.
I can easily see AI users themselves getting very soft to copyright concerns, altogether, especially the more they’re pushed into a corner with excuses like cost-saving and labour reduction. To production artists, the attraction is so, so easy to understand. Like to them, are all clouds equal? AI can render every type of cloud and simulate all styles well anyway, so wouldn’t I just be saving time and (my client) money by letting AI do it? Why stop there, why not just go directly to the source of another artist and cut and paste some of their clouds into my own work? That’s not any more shady, right? And since I’m already comfortable with clouds, I might as well directly extract some foliage and flowers for my pieces. Nothing wrong with that – who’s going to notice? It will probably just look like more AI-generated assets anyway. All eyes are equal now, right? I might as well directly copy some eyes and just distort them slightly so I’m “not breaking any rules”.
Some have argued that it can provide a starting point to paint from, like for creating background under painting, set colour palette, et al. There is that, but if you don’t have strong illustration skills and the ability to match style with what you’re generating, the results might wind up looking crudely integrated. I think there’s a subculture of photo-bashing concept artists that are naturally more comfortable dodging the grey areas of using borrowed image assets to generate pieces. The practice made its own splash regarding ethical use, but it’s a legitimate craft and some do it rather well. To artists whose philosophy is anything goes, they’re no doubt going to love what they can do with AI. But in the end, I think they’ll find there’s more control and potential in doing things the right way from the start.
It’s understandable if anyone would want to take the tech for a spin just to see what all the fuss is about, and seeing how it generates art might teach them a thing or two about how they could approach making similar subjects and styles. Just do the world a favour and don’t release your AI experiments into social media, or at least help the art world out by not putting it up anywhere that isn’t specifically designated as a pro-AI area.
Final words: make art responsibly. Ask yourself if your methods are good or bad and try to do as little harm as possible without censoring art itself. By no means will I endorse AI’s use even an inch, but if you must, ask yourself how much of an image can I responsibly call my own? How much should I alter something before it’s okay and something else entirely?
I, myself, have a long-time friend who is pro-AI but he’s all into gadgetry and loves a quick way to supposedly make a cheap buck with crypto whatever, so he was all over AI and I’m sure there’s no talking him out of it. I don’t think he’s given a second’s thought to the consequences of action. But eventually, the novelty will wear off and he’ll realize what he’s doing is cheap and cheating himself out of something he could sell himself on by legitimate means. People will repackage something as their own if they’re either conceited or desperate for $, and unfortunately, we have a lot of that going around in the world. My friend uses the excuse that he has disabilities (and hasn’t taken the long road of practice) to get good. Should I reward his effort over authentic artists conducting themselves legitimately? Even though he’s my friend, I can’t with clear conscience endorse his pieces or even offer him a seat at the table when I know he didn’t do the work others were willing to do. Am I being cold? No, I’m being fair. Because for everyone disabled that decided to do it wrong, there’s someone disabled that owned up to doing it right, and I’m not going to play favourites. How many artists are pushing through their carpal tunnel, their arthritis? Often, these really are labours of love. And people forget that. I gotta back the real artists, not the tech junkies.
Arts Artists Artwork Comments on AI and AI Art
Artificial Intelligence is here to stay, of that there is no doubt. AI has been a driving force in the social media world, search engines, modelling and has made leaps and bounds with writing, chat systems, and amazingly, with images. As a small operation looking to grow, AI has proven very useful in generating written content which can then be edited and tweaked for the purpose of educating our audience. To be honest, I didn’t see any of this coming but it has made a big splash.
Initially, the first AI Art that I saw generated did not strike me as causing any issues when creators used it to visualize new forms of architecture, dresses and other similar images. When it became a surprise was when actual known characters such as the superheroes were created with AI to portray them as babies, or combining two characters to create a new character and so forth. And while I haven’t seen direct examples yet, there is also the AI Art that has been created to portray new artworks in specific artists styles by asking the AI generator to mimic or copy the style of the specific artist.
Artists Deserve to be Paid
Since the AI datasets were created by having bots crawl the web and collect images from all over the world, billions of images have been added and a significant number of those images are considered copyrighted. Midjourney CEO admitted that this was done without securing permission to do so and confessed that it made no sense to do so, in part due to it being a herculean task. This is where a lot of the fuss comes from because AI Art cannot be created without the raw data that is already out there and produced by many artists and other creators. Take away their artwork and AI has nothing to work from, so why should AI Art sites benefit financially and not the artists and creators? It’s a very valid point.
It is also detrimental to artists and creators to take their images offline. Many sell their art and photographs in online galleries, producing prints, and other products that incorporate art. They need to have a web presence to showcase their portfolio of work. It’s the bread and butter of an artist to generate new sales opportunities through commissions, prints and original artworks to be sold.
A Solid Example of Artists Being Paid When Their Art is Used
An example I thought of is musicians and other musical artists. They write music, they record music and it gets widely distributed. Yes, some of it is pirated and given away, but there is a positive spin on that market; it results in sales. People eventually buy what they love. Radio Stations pay to play their music, too. The artists get paid incrementally for each play of their song. If the black market is popular, it feeds the radio play popularity, too. The artist gets paid. Singers and bands also decide to play songwriters’ songs. In fact, popular songs it is many artists want to sing their songs. The musical art doesn’t have to be hidden, because there is a system in place to ensure the musical artist gets paid.
Conclusion and Your Comments, Please
The future is very interesting with the advances of artificial intelligence and the creation of art, conducting research, writing new texts and so forth. Ethics are an important consideration in their use and development. Add your voice to the conversation by adding your comment below and tell us how you view AI Art, how you view the ethics of its development and the ethics of its use.
One last question: Now that AI Generates Images, Can It Feed Itself Images, too, from all these AI Artists?