The article below contains some of my personal opinions on a (currently) sensitive topic, which not everybody may agree with.
What are A.I generated images?
In 2022 we have a seen a significant rise in the number and popularity of A.I art generators. These are systems that use an algorithm to create images based on a prompt – text in which a user describes exactly what they want to see. A prompt can be one or two words, or it can be lines and lines of words describing in extreme detail the scene, style, quality, composition, color scheme, lightning, perspective etc. of the desired image. The generator will then in a matter of seconds or minutes create an image based on its database gathered from the internet, to match the description.
This new technology allows non-artists to get a visual representation of their idea, without having to hire an actual artist to create the artwork.
Some of the most popular A.I generators
I have tested some of the most popular generators and listed them below for anybody interested:
The most effective one, in my opinion, is Midjourney, which seems to have the best sense of aesthetic overall and the highest quality of images. The end results at first glance often look like works of a professional, and could be mistaken for a handmade artwork. The rest of the A.Is seemed to lack in that aspect and the images produced by them looked amateur and not very appealing.
The controversy – Protests in the art community
Recently, there have been more and more protests in the art community regarding the usage of A.I image generators. The technology has created many problems for artists who find their art has been used by these programs to learn and evolve, without their consent. A.I generated art has overflown popular portfolio websites such as ArtStation. It is also posing a threat to artists who may have less work available to them since A.I generators are much cheaper to use than a real person.
As a result, we have seen protests all over the internet by artists who do not wish for their hard work to be used in such way and without any compensation, or who are just plainly against the technology (not all though). Indeed, the companies who develop the A.Is use copyrighted images to train their generators and profit from subscriptions or one time payments made by their users. In this case, obviously, there needs to be some kind of regulation which will require these technology startups to acquire consent from artists before their artwork is being used.
Artists do reference and get inspiration from each other without necessarily asking each other for permission. However, a human mind will ALWAYS filter the information received and alter it depending on mood and skill. It will also add to it from their own experience and artistic preference. Artificial intelligence however, will process the information as is, without adding any more value to it or making it truly its own creation. As such, it shouldn’t really own the copyright to the images it produces – it should acquire it from the artists who produced the artworks in the first place.
Possible merits
At the moment of writing this article, there are many conflicts concerning this new A.I technology. I, personally think that it has enormous potential and will be widely used in the future. Obviously, a lot of work needs to be done first. Artist’s rights need to be protected in the process, before the technology is acceptable to use in any commercial work. Here are some of the potential uses for A.I generated images in the art industry:
Mood boards – generated images could be used to quickly assemble a mood board to serve as inspiration or quick idea visualization followed by the manual creation of the final artwork or concept art.
For non-visual artists – Quick creation of images to use in blog posts, book covers, album covers etc.
Texture generation – A.I could be used to generate images to be used as textures in a digital artwork. (Artists often mix in some photography to add realistic texture to their work).
To be honest, as an illustrator, I feel very positive towards A.I technology (it’s not only the tech bros!) as long as there are no copyright issues and art theft involved. I hope we see a community of artists who will like to contribute (for a fee!), their work to the A.I training programs enabling the technology to evolve in the right way. I would also like A.I generated work to be properly tagged as such, so that anybody can easily identify it.
Afterword
I don’t believe that highly skilled professional artists will loose work due to A.I usage, as their work far surpasses what an artificial intelligence can create. However, less skilled, beginning artists may have a harder time finding jobs. This is true especially on freelance websites such as Upwork or Guru, where cheap work is often most desired by clients. A.I will likely replace more and more low paid jobs, as the technology progresses and becomes better and cheaper than an average human worker. Everybody needs to be aware of this risk as they choose their profession. Making yourself irreplaceable is the biggest and most important challenge in creating a successful career.
