(3-6 December 2022) In 2 years, we have experienced an explosion in art made by artificial intelligence. Art was thought to be one of the hardest things to automate, but in a seemingly mocking manner it was the first to fall. I am uncertain and worried about what this entails. AI is a powerful tool; the automation of the mind, possibly the current century's greatest invention. The body was already automated in the industrial revolution. Perhaps this is a new one.
But I worry: man's greatest asset is his mind, his ingenuity. With man's monopoly on intelligence stripped away by machines that can outperform his mind as much as a crane can outperform his body, why should companies employ him? And whatever could he do to combat an AI should it go rogue?
It makes our current political and cultural arguments seem insignificant. There are people saying that artists will adapt to AI just as they adapted to the camera. How? An AI can outpace any artist, in whatever adaptation art makes, as long as its dataset is good enough. Artistry could just be replaced by trying to write the best text prompt. There will be no editing the generated image after, since you could just append what you want to edit to the prompt. But if for some reason you are unable to translate what you want done to your synthetic image to the AI, you will have to edit it yourself. You will not commission an artist to edit it because if you couldn't parse your request to the AI, you won't be able to parse it to a human either. The greatest challenge for those who write text prompts will be human language and the corruption that occurs when translating from mind to word. Therefore, this corruption will eventually be removed by a new technology, one that will also have great usage in destroying the human right to privacy. I think this mind-reading device will have a lot of its marketing be about its abilities to allow you to generate superior synthetic media over what you would generate via text prompt.
This is an edit from a few days after I wrote the above paragraph. I have just discovered there is a tool called InvokeAI, which is a raster graphics editor that allows you to generate synthetic images via text prompt, image to image, etc. and then combine them on the canvas and do what you please with them. Programs like InvokeAI or ones that incorporate its AI generation abilities will be common in the near future. Photoshop, Unity, Adobe Premiere—these programs, and ones like them, will likely add AI-generation features in the near future. In fact, TikTok has already introduced a feature to add AI-generated backdrops. Creativity will rise as industrial output.
Art will be killed for the artist, but infinite possibilities will be given to the layman. I am no artist, but regularly draw crude sketches of the appearances of things in the stories I dream up. With AI, I could generate images that would help me visualise my world. However, here is my problem: my worlds are primarily in the form of writing; AI will not only kill the jobs of artists, but also the jobs of every creator of media, and then some. If the case is that people will choose synthetic media over man-made media, it makes talent a worthless thing to pursue if you are wishing to make a product out of it, and so, despite knowing full-well that they are outclassed by machines, people will continue to learn to write or paint, albeit for purposes other than money.
I believe media made by AI should be split off from media made my people and given less legitimacy. New terms should be coined, such as synthetic media, meaning media created by AI. I think man-made media will be advertised as such and people will see synthetic media as soulless but useful. People might generate a funny video and share it to some friends, but they will simultaneously prefer to watch man-made media over synthetic media. When the tools are released to the public in the near future, people will start to make things never seen as possible with AI, and people whose works were rejected by companies because they were unprofitable will be able to make them anyway. A few of these things will go viral at first—there will definitely be a video game with the gimmick of being entirely AI-generated—but the novelty will eventually wear off.
Choose-your-own-adventure will become a lot better and more popular with AI. Currently it is considered more of a niche book genre but with AI it will be able to be a lot more flexible than it is in books. Using an AI chatbot, I have had my own text-adventures with it and it is indeed very entertaining. Synthetic images and synthetic videos would compliment a choose-your-own-adventure very much, but unfortunately the current image-generating AI is terrible at context and video-generating AI is, as of yet, non-existent to the public, I think.
When I generate a new image, I often examine it in awe. Yes, the images almost always fail at least one detail in the prompt, but to be able to have quality artwork can be delivered to you for free and in seconds is revolutionary. I have tested 2 models: DALL-E 2 and Stable Diffusion. The former was a lot cloudier, but I felt it was more accurate to what I wanted. The latter was sharp even though that's like pointing a magnifying glass to see imperfections. But, in terms of realism, Stable Diffusion is better. I will still use both and pick whichever gave me the best result.
Human programmers might adapt into seeing it as more of an art form, something to enjoy doing. They might begin programming in esoteric languages more and more, and might introduce constraints to their projects just for the fun of it.
If any type of media is untouched by AI, its saviour will be the inability of the client to communicate what they want properly.
I have heard of a new concept where you generate an image, select the best one, and generate another image from the previous best result, repeating and simulating evolution with a synthetic image generator. I'll do this, and will make an evolutionary tree of my own. In the video I heard about it in, the narrator compares it to exploring in a ship, and instead of the sea you're in "latent space", which is the theoretical space where every conceivable image lies. Such a novel concept, I wish I had come up with it. I am always in a rush to discover as many new ideas as possible. I get frustrated when I learn of something original that I could've discovered but didn't. I want to be a source of originality.

(14-15 October 2022) There are people saying that art will adapt like it had done when the camera rolled around. How? An AI can outpace any artist, in whatever adaptation art makes, as long as its dataset is good enough. Meanwhile a camera cannot photograph fiction or apply different art styles to images (yet). Art should have its definition changed to being exclusively of the human mind, and so should culture as a whole. These are human things and AI should not be allowed to imitate these with any validity. You have to understand that the camera is a tool, while AI seeks to combine the flawlessness and speed of a computer with the flexibility of the brain, creating an ultimate imitation of the human brain. An AI can be trained to do whatever a human can, better than any human can. In the future, you could design a musical hit in a computer, destined for success. As long as you conceal the fact that this was not made by a human, you would have success. If people discover the song is artificial, they will feel cheated and the song will feel soulless, and there would be a scandal. There will, however, be a group of “hedonists”, who don't care about if a product has human soul, and only care about how superior that product is. Should the majority be of this hedonistic group, which I actually unfortunately believe will happen, that would draw the ire of many more “sophisticated” (these hedonists would call them “stuck-up” and “narcissistic”) people, who would see it like how many now see everyone walking around looking down at their phone like zombies: soulless and almost bug-like. Who knows, however: maybe I'll be one of these hedonists if I live to see it. Hope not.