AI Generated Art - Is it contraversial technology?
The internet has been flooded with incredible images created using Artificial Intelligence drawing/art software and apps. Text to art image generation is an exciting new innovation that requires very little artistic talent, just type in a description and the algorithm gets to work creating your masterpiece. You can also input an image to "seed" the algorithm.
Now, I am no technology nerd. I don't understand exactly how it works, but I am very interested in the current discussion surrounding copyright and the general rammifications for the art community.
The discussion is hot around copyright for the technology behind this amazing phenomenom. As I understand it, for AI to learn to create images from text, the developers have fed millions of images into the technology and AI "learns" to associate text with images it has in its vast data base. The more people "play" with it, the more it learns. The latest release of a more advanced algorithm that can produce life-like human images is evenmore contraversial.
The image below was created on Artbreeder - Which allows users to create lifelike images from "genes" , or from photographs; use multiple user generated images to remix an image and "spawn" new images. I used a selfie and evolved some fascinating fantasy and ultra realistic faces. You can also create 'art' works in the style of a long list of artists and photographers.
Why is this an issue?
Here's where the contraversy begins. While is it s fun "toy" for social media and tech-savvy fun, there are some serious questions being asked.
Before I delve into this muddy water, I need to say that I have used AI to create or develop my digital art. I do not claim to have created something if I used an image that is not my own. I do often source images from Unsplash, which has a clear copyright policy. I prefer to use my own photographs, selfies and physical art as source images.
Current concerns and discussions
Copyright - AI database images
Copyright - Source images
Who owns these images?
What about professional artists and photographers? Employment future and ownership of their intellectual property.
Can a person, who typed in some words and watched AI create their masterpiece, say they did the drawing? (A philosophical question as to 'what is art?').
Art is a diverse range of human activity, and resulting product, that involves creative or imaginative talent expressive of technical proficiency, beauty, emotional power, or conceptual ideas.[1][2][3] (Wikipedia definition)
Copyright - AI database images
The discussion here is surrounding how and where developers source the images for AI to 'learn'. They state that all images were creative commons, or open source images. The sheer volume of images required for AI to properly associate text with images is mind boggling. This leaves the question open to copyright and also image ownership. If you care to go there, philosophically, then NFTs may also fall into this very grey area.
Copyright - Source images
My thoughts are to proceed with caution. Some articles suggest the end user may be open for future copyright infringements. The developers are also under scrutiny. I use only my own images, or original art, and the occasional image that is open source with attribution to the photographer.
To make a comparison; the music content creators use in their social media posts is under heavy infringement scrutiny. Some have even said that they accidentally had the radio or TV playing music in the background and had their content blocked because of an infringement. The larger platforms are very strict, so this may be a solution, but I would think it would be next to impossible to compare these generated images with the original given the user can type "In the style of Van Gough " as the prompt. Asking for a landscape photograph, hyperdetailed etc in the style of a famous photographer means AI will reference the photographer and copy it. (AI cannot reason right from wrong).
Who owns these images?
This question is way above my paygrade. If you read the fine print on some of the more popular applications using this technology, there are some vague, confusing conditions and statements that do not clarify ownership. This is more for commercial users who may want to use the AI 'people' or other scenes for advertising etc. The hobbyist who is playing with it and wants to claim the painted something themselves is less of a problem, except possibly ethically. But we do live in the digital age and we are all aware that the media are renown for manipulating photographs. You can see where this is a minefield to negotiate.
As I understand it, the fine print infers that you may not own what you create. So that, then opens the pandora's box of NFTs. I am definitely not going there.
What about professional artists and photographers? Employment future and ownership of their intellectual property.
This is a legitmate discussion, especially for illustrators and commercial artists. I guess I will go with the 'watch this space' school of thought.
One could draw parallels with the music industry and also with the introduction of new technology throughout history.
The invention of photography may have challenged artists who made a living from portraiture or scenic art. Movie then came along and years later faced competition from the introduction of videos and DVDs. Then along came digital streaming, which was predicted to cause the demise of the local movie theatre. The result was innovation in the industry. Theatres created a more immersive and customer oriented experience, and patrons continue to flock to theatres for new releases.
In the music industry there was fear and trepidation that electronic music would soon replace musicians. The drum machine is much maligned by musicians, but it has found its place. Everywhere we go we see people with headphones in, listening to music and people are still going to concerts and musicians are still employed at the local pub on a Saturday night.
T
he parallel I draw with AI art and music as far as copyright, intellectual property etc, is that the music industry seems to have come to a compromise and is actively pursuing infringements. This is another muddy area when it comes to sampling, computer generated music (similar to AI with a little more user input) and mashups. Apps that wipe vocals from tracks to produce karaoke and software that will manipulate original compositions or illegally pirate albums are easily downloaded and used.
The conclusion is that technology is far outstripping the creator's rights and legal protection for their creations.
Can a person, who typed in some words and watched AI create their masterpiece, say they did the drawing? (A philosophical question as to 'what is art?').
I certainly can't answer that question, except to put the idea forward that some would claim that text-to-art meanst that they inspired the end result from their words and imagination. If I type in "cyborg pianist" and use an image of myself playing the piano, then AI will create an image referencing the database to understand piano, playing piano etc. Then it was my inspiration. The grey area is using modifiers such as 'in the style of' an artist or a photographer.
If I post my image on social media and say "look what I created"; and not say that I used AI to actually produce the image, then where do I stand? Is a random image that AI produced actually art?
The image above was generated from text : pepperoni in space. It was a spoof on a fake photo from a French scientist. All I contributed was the concept. AI did the rest.
I would be interested in hearing your thoughts.
Conclusion
It is impossible to reach a conclusion on this subject. Most commentaries end with watch this space. As in so many other areas where innovation and technology is far outstripping the current laws, there are dozens, if not hundreds, of grey areas.
My conclusion is to proceed with caution.
Use only images you own.
Give credit where credit is due (where you have permission to use an image).
Enjoy and explore innovation. Without that we would have no Star Wars, dance beats, animated movies, funny goat reels or selfies to share and amuse ourselves with.
Resources and references:
https://www.flickr.com/people/annabuxtonart/
https://en.wikipedia.org/wiki/Art
https://www.insider.com/landscapes-you-wont-believe-are-real-2020-7#surprised-by-a-full-moon-at-goblin-valley-by-luke-simpson-1
https://creator.nightcafe.studio/create
https://deepdreamgenerator.com/
https://www.artbreeder.com/beta/browse
https://petapixel.com/2022/08/19/ai-image-generators-could-be-the-next-frontier-of-photo-copyright-theft/
https://labs.openai.com/
https://unsplash.com/
https://help.unsplash.com/en/articles/2612315-can-i-use-unsplash-photos-for-personal-or-commercial-projects
https://opensea.io/explore-collections
https://creator.nightcafe.studio/creation/9CggD9eMDhKhB80PTfGy