I have wanted to explore AI image generation for a while and finally got around to it tonight in front of the fireplace and with the neighboring café playing live jazz.
Here is one of my first experiments with Midjourney. A neo-shaman in Tokyo in the rain with dramatic backlighting. I love that he or she is covered in plants and flowers.
I have seen some discussions about AI-generated images.
CONCERNS ABOUT NEW TECHNOLOGY
Will it replace human artists? Will it make it possible for people to make their own illustrations instead of commissioning photographers and artists? Will it ruin creativity?
Yes, some of that will probably happen.
And it’s also important the remember that these are the type of concerns that predictably come up when new technology comes onto the scene. And each time, the new technology finds its place among everything that has existed before and continues to exist.
When photography came, people said it was the end of painting. What happened was that it caused painting to change. Much of it became more free, imaginative, and abstract, and photography and painting not only co-exist but inspire each other. When CGI became viable, people said it would replace practical effects and even actors. In reality, CGI co-exists with practical effects, and it has even led to new types of jobs for actors in the form of motion capture.
I assume something similar will happen now. Some will use AI for illustrations. Some will continue to hire artists and photographers. AI art will inspire human-created art. Human-created art will continue to inform AI art.
It’s not either-or, it’s both-and. And it may well be that the interplay between AI and human visuals will create a kind of artistic and creative mini-revolution.
It’s also very likely that human-created art will be valued even more. AI art will make it more prestigious.
CULTURE MEANS LEARNING FROM OTHERS
Some say that AI steals people’s work to create new work and make money on it.
I understand that argument and concern.
And I also know that that’s culture. That’s what people have done from the beginning. We learn and take good ideas from each other and do something different with it. That’s how we have a culture in the first place.
The AI is just a bit more comprehensive and effective than any human can be, and also a little less creative.
WHO DO THE IMAGES BELONG TO?
Another question is: who owns the images?
In a practical sense, it’s determined by the AI companies and the law.
And in a larger sense, they come from the collective experience and creativity of humanity and really from the whole of existence. It’s always that way, no matter which particular human or technology it comes through. It’s just a little more obvious with AI images.
Some also criticize AI-generated images because they reflect cultural biases. They learn from our culture so they will inevitably reflect biases in our culture.
For instance, if I don’t specify ethnicity for a portrait, I get a European person. If I ask for a god, even a traditional Hindu god, I get someone absurdly muscular. If I ask for Jesus or his parents, I get Europeans and not middle eastern people. If I ask for a general person, I get someone unusually good-looking in a conventional sense
I would say that’s equally much an upside since it brings cultural biases – picked up by and reflected back to us by the AI – more to the foreground. This leads to awareness and discussions – in the media and among those exploring AI art and the ones they share these reflections and observations with.
A lot of people are more aware of these kinds of cultural biases now because of these AI images.
MY OWN BIAS
I have a background in programming and in art, so I naturally love AI-generated visuals. I see it as a way for people without too much experience to still create amazing images. It’s a way to generate ideas. And it has its place and will co-exist with old-fashioned human skills and creativity.
UPDATE AFTER ONE WEEK
I have explored Midjourney and AI image generation for a week now, and find it seems to fit me well. It’s fun to see images created that I have had in my mind for a while but haven’t created in pencil or oil. It’s also fun to get to know the AI and sometimes be surprised by results better and more interesting than I imagined.
I also find I cannot really take ownership of the images, apart from in the most limited sense. They are generated by the AI, the AI is trained on perhaps millions of images created by others, and it’s really all the local products of the whole of existence – going back to the beginning of the universe and stretching out to the widest extent of the universe (if there is any beginning or edge). It’s always that way, and it’s even more obvious with AI-generated images.
The images are very much co-created by me, Midjourney, innumerable artists whose works have informed the AI, and all of existence.
I have also started an Instagram account for my AI image experiments.
Note: Specific prompt for the image above -> Neo-druid shaman in Tokyo 2300 rain dramatic colorful backlighting semi-realistic