I’ve been experimenting with AI image generating tools recently. What I saw starting with random weird and too brightly colored collages generated from simple texts like ¨pink pig swimming in a yellow lake¨ last year, has evolved to very complex detailed imagery created using not less detailed prompts like ¨pink happy looking middle sized pig swimming in a yellow lake with big waves in Hokusai style¨ Just an example. (I didn’t try with this prompt yet but why not? You’re free to use this one and maybe you can add some Naboo renaissance architecture in the background)

There is a great number of text to image AI tools, such as Midjourney, Stable Diffusion or DALL-E but of course most of them are priced, like 10 dollars a month in the cheapest version. so I headed to some quick free options just to have an overall idea about AI image generation.

Of course free tools like Picsart produce less impressive images than its high priced counterparts and this one is a not-so-enchanting example. You will notice that it can’t foresee (yet) that an iStock logo isn’t something that I or many other people would absolutely love to include in an ¨art¨ that we generate.

Of if we are lucky enough and with the right prompt we can get better looking images like this one

The prompt was ¨Istanbul Bosphorus bridge sunset in the background.¨ At times I added 1800s colored lithographic style or ukiyo e style and I opted for anime or manga styles

Paid tools like Wepik come with more beautiful looking images with a similar prompt. And with more artistic flair by including a semi-psychadelic and fractalistic version of a Tatooine sunset.

But all the tools I tried were unable to include the very Bosphorus bridge that actually look like this

Wepik was even more impressive when I tried to generate floral ornaments in Victorian or maybe Art Nouveau style and it didn’t forget to include some hot girls I didn’t ask for

And everytime I included ¨no people no girls no women or men or any people in general no human faces no bodies only ornaments only flowers or leaves or whatever but only ornaments¨ in my prompt it generated beautiful girls beautiful enough to make me feel more insecure about the way I look. I always knew I wasn’t that close to Charlize Theron in appearance and I was ok with that but stop ai really.

So yes, I was impressed and also frightened. As an artist who enjoys creativity and who always seek for originality and novelty in the works I’m, for a not very short time, constanty thinking about the future of art with AI. I already see people who are labeled as ¨AI artists¨ selling works with the help of AI generation. I know some artists and illustrations filed lawsuits. You can read the details about the artist Kelly McKernan, Sarah Andersen and Karla Ortiz’s lawsuit against Stabilty AI, Midjourney and Deviantart who are responsible for creating or using AI text to image generator Stable Diffusion.

Everytime we put out an artwork on the internet it’s highly likely that it will be included in the database that these image generators use and it’s as much likely that when some people generate art they will get an artwork that includes some elements of our works. Call it an inspiration or stealing but it will be in most cases without our consent. Artists include their names in prompts like ¨this and that and this and that in the style of (artist’s own name)¨ and are terrified to see how AI literally stole their work. Concerns arise about the lessening need for artists and illustrators and the loss of jobs. So will AI replace artists, the world is asking.

The question of replacement: To what extent AI can replace the organic human brain?

AI creates beautiful images. At many times boringly beautiful. Too bright colored childish woo woo in my opinion. Most of them are better suited for spiritual websites about magic, tarot and loving caring 5D aliens who will watch over us until we ¨ascend¨ Yes those type of images. It can copy, synthesize and collage pieces from any style you want actually, not just woo woos. But that’s all it does. It can’t put a soul to it, it can’t generate emotions. It’s robotic. It can know meanings but can’t find out the context behind these meanings and do a proper synthesis out of them.. To better explain I’ll give examples.

Take the logos by Herb Lubalin. He created Marriage logo by making the two Rs like a married couple. Letters in Families logo look indeed like a family and the ampersand in Mother and Child logo is placed in the O of Mother so that it looks like a child in the mother’s womb that is the O. Can AI do a similar job really?

Isn’t it a place where a need for human organic artist comes in? I know many other examples that comes from the same mindset and I try do apply this mindset in some of my lettering works too.

I still don’t know about the future of AI and don’t know if AI will be that smart to be able to unite shapes and meanings in one pack and reconstruct a shape to convey a meaning that is actually unrelated to the shape, I really don’t know, we’ll just wait an see and then we’ll have a lot of more reasons to be intimidated by the AI (and this without mentioning apocalyptic theories on what AI may intend to do in the future) but for now I can say artists who are at the highest level of creativity are in a safer place. Especially if the employers are conscious enough to hire human artists and care to include elements of the human touch in the works they present to the public, be it book covers or magazine illustrations or comics. And in general artists may be forced to focus in their artworks on what only they can do and AI can’t do, if they choose art as a career path. So for the careers in the digital arts upskilling will be the key. But I know digital art didn’t kill traditional art so we can always take our brushes and happily spread our oil paints on our canvases and there will be a demand for it too.

Indeed there is something that I would love to see AI doing. It’s to scan my brain with some microchips and detect the flashes of images, especially illustrations and ornaments that briefly appears in my mind and disappears, record those images and send it to my computer as jpeg or pdf or vector files if possible. I see those images at night in my bed before I sleep and I try to memorize some of them to draw and paint in the next day but a technology that will record these would be extremely helpful. Unless it includes those images in a database where other AIs can extract and include them in their own works of course. Anyway forget it.