In my previous blog posts, I've explored whether artificial intelligence-generated images qualify as art and how they might impact artists by potentially taking away jobs. Despite my concerns, I remain optimistic about art as a hobby. However, there's a real possibility that AI-generated images could displace some commercial artists; of course, this potential negative impact of AI extends beyond just the art world. In this post, I want to discuss a potentially positive outcome of AI for classic artists.
Now, there's a whole can of worms I don't want to get into regarding AI images and copyright, specifically concerning AI-generated images potentially infringing on someone's copyright. But if an AI-generated image is broad enough not to mimic someone else's work, it essentially holds no copyright and becomes public domain. This opens intriguing possibilities for artists seeking reference photos. Many artists know the challenge of finding a usable image without infringing on copyright. If an artist significantly alters a found photo, they might avoid legal issues, but achieving photo-realism while steering clear of copyright infringement is tricky.
Consider the utility of AI in generating public domain reference photos. Imagine needing a reference for a drawing of a mountain lion by a stream. Options include drawing from memory, piecing together various photos, risking copyright infringement by using internet-found images, or using an AI image generator like Midjourney to create a unique, copyright-free reference with a prompt like:
"Majestic mountain lion, fur glistening with droplets, standing in a crystal-clear mountain stream surrounded by lush greenery and rugged terrain, early morning light casting soft shadows --ar 3:2"
Here are the initial photo options the AI produces:
If you did this yourself, you would receive four different images. Running it 100 times, each image would slightly differ. You can make modifications or variations. It's not identical to the creative process involved in making an actual piece of art, but in terms of selecting a reference photo—which might be as simple as conducting a Google search—this approach involves a more engaged creative process. And you can rest assured that you're not infringing on someone else's copyright because the photo is general enough to be considered legitimately public domain. Then, if you infuse your own style into the final piece, the art becomes as much your own as any creation derived from a reference photo.
Now, if you were to take the photo yourself, such as capturing a mountain lion in a stream, that would be even more uniquely your creation because you're both the photographer and the artist. However, not everyone has the opportunity or ability to do so, making this a less than perfect, yet viable, way to obtain reference photos for artists.
What if you need a very specific and sort of odd reference photo? Like a hippopotamus standing in Times Square, for example. Searching for a "hippopotamus standing in Time Square" yields no relevant results on Google. However, an AI prompt can produce a highly specific and unique reference image.
“Photo of a realistic hippopotamus, standing in the middle of Times Square --ar 3:2”
The AI is taking over a bit more of the creative process, as some of the creative choices for the reference photo depend more on the algorithm than on you. However, every artist who starts with a reference photo that isn't their own is outsourcing some part of the creative process to someone else. At least this way, they're not ripping off another person’s work.
Now, let's say you want to create something surreal, like an apple melting on a table. There are many examples of this kind of image available through Google, but you're looking for something more original to respect other people's copyrights. So, you input a prompt into Midjourney like "Photo of a melting apple --ar 3:2."
With a bit of tweaking, you'll have a photo reference for a surreal drawing or painting. But in surrealism, it might be better to rely more on memory or to combine different photos. Strictly copying a reference photo, even one generated by AI, may dilute the concept of surrealism, which is meant to represent dreams and your own creativity. However, if you find this method the best way to achieve your desired results, who's to say you're wrong?
Yes, I still harbor significant concerns about AI and its potential impact on artists, particularly those in commercial fields. Despite striving to maintain an open mind, I find it challenging to regard AI art as a genuine form of artistic expression. However, as a tool for generating reference photos for artists who draw and paint by hand, AI could offer substantial benefits. To leverage this, one would need to familiarize themselves with using AI image generators, and fortunately, there's ample information available online for this purpose. Once you generate an image to draw from, you can enlarge that photo with your LUCY Drawing Tool, sketch it out on your canvas, and thus create beautiful art. Isn't that ultimately the point?