AI: The Djinn Is Out Of The Bottle

AI: The Djinn Is Out Of The Bottle

AI. It’s a thing that has been out there for longer than you think. Have you ever talked to Siri or Google? Have you ever relied on a spelling or grammar checker like ProWritingAid? A good deal of us use these things without a second thought. But, when the visual AI came onto the scene, it divided many people. I have found two predominant die hard schools of thought on AI have emerged. The first being the group who really thinks they are doing no harm and that the output is NEW and the machine LEARNS. The second group firmly hates AI and declines to support its use. I just don’t think this issue is as black and white as we want to make it out to be.

I hate that I have friends in both groups and they seem to fight a little like Capulets VS Montagues. And I hate that I feel somewhere in the middle even though there seems to be no amicable middle ground. When I was on the AI panel at Imaginarium, I learned many people out there believe the use of AI will all boil down to ethics. Now, I’m sorry, but humans don’t have the best track record with ethics. There will always be those who toe the line and use proper artists and seek proper licenses. And there will always be those who don’t have a problem doing whatever the hell they want—such as these companies who made the programs. They’ve unleashed a Djinn upon the world and it’s not going back in the bottle.

As I sit here and watch the advancements in AI, I know I cannot compete with the speed and ease of creating art like the machine can. Yes, it still adds weird eyes and 47 fingers to things, but some people will overlook that just to save time and money. Remember the old saying if it’s fast and cheap, it can’t be good. Which leads me to a grey area where stones are often flung: You can use AI for some things right now with minimal damage to human creators.

Problem: AI needs to be ethically sourced. This is my biggest problem with visual based AI and even word based AI. These companies just went and stole a bunch of stuff on the net to “train” their programs. They didn’t ask. They didn’t let people opt out. Do you remember when people first started blogging and the number one rule was you can’t just Google an image and use it for your website? Or how about the “artists” who straight up steal other people’s art and use it in their finished works? I think that school of thought to not engage in these kinds of practices is still supported by most humans.

Solution: Let artists opt in or opt out. Let artists who WANT to contribute do so and compensate them. Just like those artists who make stock photos and assets. These tech companies are MAKING BANK on these AI programs. They can afford to pay at least as well as a stock site or even Spotify for the use of the art.

Problem: People are selling AI images and making money. This is wrong. Not because it is AI, but because of the above problem where the “learning” was based on stolen goods. They say even the written AI modules are out there stealing things to “learn”. It’s not cool when you make an entire book and cover using just AI. It cheapens the entire creative process. Writing a book and making art both hard endeavors which require years of REAL learning. It really sucks someone can just suck the stuff off the internet, click a button, and make a ton of money with only like a day of half effort.

Solution: Until the dust settles, or a company comes out with an ethical model, AI shouldn’t be used as the main focal point of anything for sale. Not your book covers. Not your merchandise. And for the love of all things holy, not your entire freaking book. This kind of use HURTS the human creators. It saturates the market with shit and reminds me of companies who make cheap knockoffs. There is no heart or passion in the knockoff. And CREATING requires more than simply MAKING something. Yes, it’s great to make money on art, but that art should be something that comes from YOUR hard work and learning. Not some tech guy’s algorithm mashing stolen pictures and words together to regurgitate a product.

Problem: AI is cool as fuck and it’s here to stay, so how can I use what’s available ethically?

Solution: I think people can use AI right now as an excellent tool. A jumping off point. You can have entire conversations with Chat GPT. Some of us don’t have a human sounding board to bounce things off of. AI is good at finding a starting point when your thoughts are all jumbled up. It’s great for lists and product descriptions. AI is also a good helper for dreaded blurbs for both books and people alike. Of course, you have to refine all that stuff, but it’s a good place to start. Creating anything can be a lonely journey and not everyone has an author bestie or a tribe. But just about everyone can talk to Chat GPT.

AI is also good for communicating an idea visually to your human artist. I’ve found that it’s often hard for the non-art person to tell the art person what they see in their head for a commissioned work. What has always helped me are pictures clients send me and say this is the aesthetic, theme, etc., I am seeing. You obviously don’t copy those verbatim, but they inspire you to breathe life into an original work of art that both you and your client love. I don’t see using AI in this manner as any different from a mood board or an inspiration board you make on Pinterest. It’s a starting point a spark of inspiration.

There is also this whole issue with the Adobe Generative Fill. Adobe says they ethically sourced their images. Some artists say no, that’s not true. They have even broken ties with Adobe and they are professional artists who have used PS all their careers. I honestly don’t know who to believe. I have no idea where Adobe got their stuff and most of the arguments I have found are people confusing the generative AI with the stock photo site where people may have used another AI to make a stock photo, which is apparently not allowed by Adobe Stock at all. It’s a real crazy mess.

I do know Adobe has put out a clause that makes the generative AI safe for commercial use. And as a stickler for always reading the licenses and such, I’m leaning towards having to believe them. If this was some other technology that I stumbled across and they said it was OK for commercial use, I probably wouldn’t question it. I’d screen shot the hell out of it and go about my way. Frankly, until someone proves Adobe a liar, like they have with other AI platforms, I have to count Adobe as being truthful. I can see a million great uses for ethically sourced AI, which is what Adobe claims to have. I haven’t been given concrete proof to the contrary at this time, but I am always watching to see the AI developments.

Full transparency though, I’ve played with the PS beta, and I find it a really cool process. Again, the generative fill and new tools are cool to HELP the human artist create a great image, not make an entire picture. The one image I did make, I used the fill to help me make a background of a demon world. All the other elements and post work effects were done by me like how I would in any of my art projects. I really liked the outcome. In fact, I’ll post the image here in this post.

Only the background is generative fill. The lighting, character, rock she is standing on, and post work are all mine.

I also think AI is fine to use for social media posts, online ads, and other throw away “free” things. Now, before you throw stones at me, listen. I used to make money in this area of the art world. I also make 3D renders, which look an awful lot like what the AI finished products look like. However, like I said before, I can’t keep up with the speed that AI can spew out images. The time it takes me to make a character, render it, and then do post work? AI has already made 10K images. And let’s not even talk about how long it takes me to draw something by hand!

In today’s advertisement world, people need good art in a super fast turnaround time. Most human artists can’t do that. That’s why in the past, most people subscribed to stock images for their advertising needs. But you know what happens with stock images? Everyone can buy them and then people out there shopping get confused because all the ads seem to have the same 10 pictures because those were the best that site had to offer. I started making stock renders of dragons to sell because book covers kept featuring the same dragon, which wasn’t even very good, because of the limited options on stock sites. Not everyone has the time or patience to get into 3D rendering, which is a complete process above and beyond photomanipulation. Trust me, 3D rendering required gigs of props, technical knowledge, and patience. 3D renders may look like AI, but they are nothing like AI.

Most consumers judge things by the “cover”. The pictures in ads, the music, the colors…it all serves to grab a potential customer’s attention to pull them into looking at what you are selling. And one surefire way to stand out is to have not only branded material, but unique graphics. I can’t compete with AI in this area, and I’m thinking I’m just going to have to suck it up and adapt because the djinn is not going back into the bottle no matter how many people refuse to engage with said djinn—and for me that area of my art sales has sailed away for over a year now. Vella covers, character concepts, ads…people are using AI for most of those now. Even people who still come to me for cover art or branding will use AI for some of their things.

The biggest solution here, besides not using AI at all, is to simply treat it as a tool. Continue to create your beautiful things. Don’t stop creating and pushing your skills. Continue to ethically produce your own creations. And if you need to hire the work to be created for your merchandise, character art, books, or your book cover, FIND A HUMAN. There are people who will write an entire book for you. It’s called ghostwriting. I don’t believe in using a ghostwriter, but many people do for books, blogs, and anything they can outsource. The people who write the words get paid, but typically get no recognition for the finished product. There are also people out there who will sell you an entire plot concept and the 12 point outline. Again, not my thing most of the time, but some people love to use those tools. And I think that’s perfectly fine! There are even hundreds of legitimate artists out there who can help you get art you will love within your budget, you just have to ask them!

In the end, I don’t believe AI is going to slow down and even with some of the newer rulings about AI art and copyright I think we are going to continue to see an increase in AI produced works. So, all I can ask is that you don’t pay these shady AI companies to give you some kind of mass produced cheap knockoff. Don’t cheapen your final creation by wrapping it in tissue paper when golden wrapping paper is easily available. Remember, fast and cheap is best for burgers and pizza delivery, not entire works of art.

What are your thoughts on Chat GPT and AI art generators? Do you think it all boils down to ethics? Are you on the side of embracing the AI trend or on the side of AI will ruin jobs? Would you be 100% on board with AI if the source material came from an ethically obtained pile of art or writing?

Please note, I will monitor the replies here and anything hateful won’t make it onto the comments.


Commenting area

  1. I’d like to be able to give an AI tool images that I created of people, places, things, etc., and have it generate images of these identifiable-as-the-same people, places, things in other poses, situations, etc. That is, I’d like an AI art tool that I can seed with my own stuff and have it produce more art from my own stuff. I don’t see how that could violate any copyright, and I wouldn’t have any ethics concerns about it, either. It would be my intellectual property that I’m manipulating. I’d just be creating more of it faster than I can without the help of AI. For instance, I’d like to feed an AI the image of a standing person, and have the AI spit out another version of that same person, but in a squatting position. I can do that with 3D render software, but it takes so much time. The 3D rendering software would make the person in the new image identifiable with the person standing in the original image, whereas AI software that I’ve seen doing something like this so far doesn’t preserve facial features to the same extent. Maybe it will get there eventually. It would then be a matter of licensing software to do AI-based tasks for the individual, as opposed to stealing art to feed the AI and then selling the output of the AI.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.