This is an AI-generated artwork of a person who doesn’t exist
This, that you see right here, is an artwork I produced using AI. First creating a person that doesn’t exist at all, with a GAN model (check https://www.thispersondoesnotexist.com/ to get one yourself!). Then, creating the painting artwork by using a styling model.
Recently, I got the unluck of running into an artwork that reminds me of the famous “Banana on the wall” piece. Only that this time, it was eluding animal abuse, actually using the hashtag for it.
The reason I came across that image was that recently I have been reminded about the huge problem in social platforms related to abusive and violent content. Especially towards animals, who sadly don’t have the same rights in all countries. Like for example, Hungary, Finland, Romania, and even some states of the United States. I consider leaving that topic about animal rights and social media, for another time.
When I saw that image I was outraged, I only thought about the millions of animals that are oppressed by our society in the dairy industry, meat industry, leisure industry with bullfights, and now in the art field.
I called out the situation, speaking my mind and opinion in a, maybe for some, exaggerated way. One of the responses I got, from one of the artist’s friends, was that I didn’t understand art, and if I considered this not to be art is because I don’t have a brain. The rest of the “arguing” in the responses was mainly about how I look, and other of my physical characteristics. They wouldn't reply whether or not they are supporting/promoting animal abuse through this type of “art”. But instead, every time I asked about it, this person only changed the topic back to my looks and content, complaining that what I do and this field is useless and no one sees value from it.
I thought about it, while I was doing endless homework for university and work. And I thought how different some people can see automation, technology, and their inevitable influence in the future.
Perhaps some people thought that algorithms could never be better than them as artists, painters, sculptors, or scriptwriters. And perhaps they are right.
Or maybe one day AI will be able to recreate a "Banana on the wall", and understand art without the need of having a human brain.
What I can be sure about, is that there is still a long long way for tech to continue surprising us in all the professional and non-professional fields known by men, today.
Try yourself!
I would love to share with you some free, online tools for you to try to generate art using AI:
- Runway ML — An easy, code-free tool that makes it simple to experiment with machine learning models in creative ways. Our overall staff pick.
- GANBreeder — Breed two images to create novel new ones using GANBreeder. (Note that GANbreeder was renamed ArtBreeder, with several AI models to manipulate photos).
- Magenta — An open-source research project exploring the role of machine learning as a tool in the creative process. (Coding skills required).
- Processing — A flexible software sketchbook and language for learning how to code within the context of the visual arts. Includes p5js (Processing for JavaScript) and Processing.py (Processing for Python). [Processing does not use AI, but is a great tool for generative visual art].
- ml5.js — ml5.js aims to make machine learning approachable for a broad audience of artists, creative coders, and students through the web.
AI-Generated Music/Sound:
- Magenta Studio — A collection of music plugins built on Magenta’s open-source tools and models.
- AI Duet — Play with a piano that responds to you.
- NSynth Sound Maker — Create your own hybrid sounds and instruments.
- MuseNet — Generate 4-minute musical compositions with 10 instruments, and combine styles from country to Mozart with MuseNet (also available on GitHub).
- Pitch Detection — Use a pre-trained pitch detection model to estimate the pitch of sound files through a computer mic.
AI-Generated Images / Pictures:
- Deep Dream Generator — Stylize your images using enhanced versions of Google Deep Dream with the Deep Dream Generator.
- DeepArt.io — Upload a photo and apply different art styles with this AI image generator, or turn a picture into an AI portrait of yourself (also check out DreamScope ).
- Visionist: Upload and apply AI Art styles to your photos, including abstract filters, cutout portraits, and more (iOS. Made by 3DTOPO Inc.).
- GoArt — Create AI photo effects that make your photos look like famous portrait paintings with this AI image generator. (Web, Android and iOS. Made by Fotor).
- Deep Angel — Automatically remove objects or people from images. (Web. Made at MIT).
- Google Deep Dream — GitHub repository for implementing Google Deep Dream.
- GANBreeder — Merge images together to create new pictures, make hybrid AI portrayals and create wild new forms that have never been seen before. (GANbreeder is now called ArtBreeder).
AI artwork sells for $432,500
The portrait in its gilt frame depicts a portly gentleman, possibly French and — to judge by his dark frockcoat and plain white collar — a man of the church. The work appears unfinished: the facial features are somewhat indistinct and there are blank areas of the canvas. Oddly, the whole composition is displaced slightly to the northwest. A label on the wall states that a sitter is a man named Edmond Belamy, but the giveaway clue as to the origins of the work is the artist’s signature at the bottom right. In cursive Gallic script it reads:
This portrait is not the product of the human mind. It was created by an algorithm defined by that algebraic formula with its many parentheses. And when it went under the hammer in the Prints & Multiples sale at Christie’s on 23–25 October, Portrait of Edmond Belamy sold for an incredible $432,500, signaling the arrival of AI art on the world auction stage.
The team collected a set of 15,000 portraits from the online art encyclopedia WikiArt, spanning the 14th to the 19th century, and fed them into the GAN algorithm. GAN algorithms have two parts: the generator and the discriminator. The generator learned the ‘rules’ of the portraits, “for example, everything has two eyes and a nose,” Caselles-Dupré says, describing a process that takes about two days. Then it starts to create new images based on those rules. Meanwhile, the discriminator’s job is to review the images and guess which are ‘real’ ones from the dataset and which are ‘fake ones’ from the generator.
It’s hard for us to naturally think that AI can be our ally, instead of our enemy. But that’s why we all should take part in this new, modern revolution.
