Artificial Intelligence (AI) imaging has exploded on the internet. So many generated images are coming out of AI imaging platforms that it is CRAZY!
Many images are rubbish but many *(hundred thousands every day) are really astonishing, and we wonder what will happen with the creative industry, artists, education and our jobs.
I have tried to play with Deep Dream Generator, Dall-e platform, Midjourney, Stable Diffusion , DreamStudio, Nightcafe and mobile app Wombo ai, and they all have text to image features and some have also image editing options.
I prefer to work with Midjourney (MJ) text to image Bot because the artworks can turn out to be lovely. Also, the MJ community on Discord is very helpful, so I quickly learned how to write my PROMPT (text instruction to bot) in order to get the best results.
But still, they all (AI bots) have some troubles with generating eyes, hands and flamingos 😀
I used the simple prompt “draw me a sheep”, from the famous little novel “little prince”, and here are some results :
The truth is you can to learn how to write the PROMPT (image description) to get better results, but even than you will be always surprised with the results.
I have a feeling that there will be a lot more of these AI imaging platforms. People are playing with creativity, generating images, visualising their dreams … and they are paying for it. Yes, although they call these platforms Open, Beta, or whatever …they all have subscriptions or premium options.
Artists are not very happy about this AI situation especially those with distinctive style, because people can drop a phrase like ‘Picasso style’ in the prompt and make their generated image look like the artwork of Picasso. The biggest problem of course is with contemporary artists and illustrators, concept artist and character designers, because AI does their style quite well.
With this AI generators, people who know to use right keywords in their prompts can be great artists. There are many Instagram accounts with AI generated feed from people who are not artists.
Artist also use these platforms so they can explore their work or deepen their style, or just to play like everyone else. I love how Vini Naso generated masks that are quite similar to his style.
Co-Creation, authorship and Copyright
So who is the author of these artworks in the end? Can you own them and sell them?
For me, this creative process with ai image generators feels like when you read a book and go to see the movie based on the book, and then you realise that they have changed some scenes and characters, and they don’t quite look like you have imagined, and sometimes it turns out to be a great movie and sometimes it sucks completely.
So that is how Co-creation with AI feels like for me, because sometimes images are not matching text prompt but they look great nevertheless, and you get thrilled and surprised but you realise that the bot has the mind of its own.
Here is example where it generated a lovely image but the prompt doesn’t quite match:
This is from Midjourney’s community showcase where you can browse images and prompts and reuse them in Discord app. That means everybody can take your prompt and reuse it. If you want to work in private, you have to choose more expensive subscription.
I read somewhere that soon we will have new jobs like: senior prompt expert, or prompt database technician or something like that.
And that is an interesting thought.
Considering the copywright issues, those platforms have their own policies and basicaly they permit printing or minting (as NFT) your ai generated images, but you have to subscribe to get valid resolution of files, or to use them after a limited trial period.
Here is an excerpt from NightCafe Studio FAQ and Midjourney Terms of Service:
Here are some of my experiments in MJ, Dall E NightCafe and Stable Diffusion:
Some are more or less successful, but when I tried to generate two flamingos they all failed. (or did they?)