I had the opportunity to explore two different AI image creation platforms. There were a few different ways to create the image; I could input a text prompt, input an image to be altered, or I could ‘work collaboratively’ with the AI. We used two platforms – Playground, https://playground.com/ and Leonardo AI https://app.leonardo.ai/ . I found Leonardo to be the more versatile of the two.
These were the images I produced through AI generation. I directed it to include ‘dappled light through trees’ and use an impressionist style. I needed to adjust the prompts many times before it produced some images that I considered fit the brief. The impressionist influence can perhaps be seen barely through the brushwork but was not evident in the colour scheme, which the AI would not make more neutral or natural.
The algorithm tended toward wide palettes of vivid colours, smooth tones and simple symmetrical compositions and I found it very resistant to changing these conventions, even when directly prompting it to. This was especially true of the ‘collaborative’ mode, where the image outcome appeared nothing like the marks I had made.
Though I acknowledge that AI can be a useful tool, I don’t feel like I would consider the above images to have been produced by me because they are not reflective of my style and I didn’t have any physical hand in their creation. If I had given the prompt to another artist to create the work, I would have felt the same. Perhaps it would be different if I apply the AI for a different purpose or in a different manner- I may explore this further in future.
AI art is a vast and complex topic and I have many different (sometimes conflicting) thoughts about it. I won’t go into detail here, but I’ll briefly note some thoughts for future exploration:
- AI could help us understand and explain what we innately know
- It doesn’t know what to do with unconventional people or anything that doesn’t fit its narrative- everything becomes a stereotype of itself
- People had concerns at the birth of photography and art didn’t die. AI is simply another tool
- Workers in lower economic countries are exploited to moderate its content when training the AI. They carry out traumatising work for below minimum wage
- It could take us to a future where capitalism increases in scale and pace, and inequalities become more pronounced
- It can emulate us but doesn’t understand us
- It makes the artist redundant, doing the work we want to do, whilst it can’t do the work we don’t want to do
- It is trained on stolen art
- Creating something the hard way doesn’t make it more valuable. Using AI as a tool doesn’t reduce it’s value
- It can only emulate, not innovate. It is formulaic, not creative. It discourages original thought, and genuine voices could be drowned out by artificial ones
- We don’t need AI art but we do need artists
- It removes the process and journey of creating the work, which is part of its value
- The AI algorithm decides what styles and themes are more valuable over others. There is no taste discretion, undervaluing the work of those who are more experienced
- Due to its vast database, everything gets homogenised. For the same reason, it has the benefit of our collective intelligence and makes this available to all