I have explored AI art for about ten days, mainly using Midjourney. (AI art is, in short, software that creates images from text based on having analyzed perhaps millions of images.)
I am not all that familiar with the discussion around this, although I have picked up a few things here and there.
Here are some additional thoughts from my side.
INTELLECTUAL PROPERTY RIGHTS
There is obviously a discussion about intellectual property related to AI-generated images.
Personally, and so far, I am just exploring it for fun and I share a few images on social media. It’s perhaps a bit similar to creating collages using other people’s works, which is what people did before AI art.
If someone makes AI art in the style of a specific artist and sells it for money, that’s more questionable.
But what if the style is a more generic one or one that cannot be pinned on any one particular artist? Is that too a problem? Some will say yes since the AI is trained on the art of many artists who unwillingly contribute to the AI result.
It also seems clear that to many, AI art is an exciting new frontier. It brings professional-grade image-making within reach to more people. And AI art may inspire human artists, just like human art informs AI art. It’s likely not possible to put the genie back into the bottle, and would we want to if we could?
For me, what it comes down to is: (a) It’s good to have these conversations. (b) This is a kind of wild west where the law has not yet caught up with the technology.
AI art is informed by images in our western and global culture, and obviously reflects biases from the material it’s trained on.
For instance… Jesus and his human parents are depicted as white Europeans, not Middle Eastern. A secretary is assumed to be female. Unless something else is specified, people come out young, white, fit, and beautiful according to western conventional standards. The default man comes out very muscular. And so on.
Some are concerned that this will reinforce existing stereotypes, and that will probably happen in some cases.
The upside is that the inherent – and quite obvious – AI bias leads to conversations among people and in the public. It makes more people more aware of these biases – the standards, norms, and expectations – in our culture.
When I was little, I loved stories about the exotic – other parts of the world, other people and cultures than my own, other landscapes than I was familiar with, science fiction, and so on.
I wonder if this is a natural fascination. We may be drawn to what we don’t know, partly because it helped our ancestors be familiar with more of the world and this aided their survival. The unfamiliar and exotic are also good projection objects, which tend to create fascination.
When I make AI art, I often find I follow my childhood draw to the exotic and unknown: Shamans, different ethnicities, fairy tales and mythology, UFOs, and so on.
Is this problematic? If we exoticize certain ethnic groups and people and think that’s how they are, then yes, to some extent. It’s out of alignment with reality and whether we idolize or vilify, it does these groups and people a disservice.
Is it inherently problematic? Perhaps not always, at least not in the sense that it harms certain groups.
For instance, I have a series of images of neo-druid shamans in the future and make sure to include a wide range of ethnicities and ages. This is a form of exoticism, and I aim at making the exoticism universal and include all types of people.