TYATautYearningArt
patreon
[ENG] Do AIs Sketch the Dreams of Doujin Artists? vol. 02.5
🕑 Added 2024-03-27 12:25:57 +0000 UTCThis article is a summary of just the introductory part because the amount of text has ballooned too much. I will write the main topic in the next article.
++++++++++++++++++++++++
My fat cat is 23 this year.
・Preface to the Preface - Japanese love to "character"
----------------------------------
First of all, I must apologize that my writing so far has been primarily for Japanese people, and I think it is really hard to read with my clumsy translation.
And today is an era of information overload, and even in Japan, things like fast movies that are summarized in about 10 minutes and catchy short sentences written by someone even though the source is unclear, are competing every day on SNS, etc., what is called Fast 〇〇 is becoming a problem. In that sense, AI illustrations are “fast paintings”. It’s natural that the value per sheet is decreasing.
First of all, how many people have read up to this paragraph? Why am I writing such a thing?
Despite the fact that “text fatigue” is said to be occurring worldwide, Japanese people are said to like text.
Currently, as a method of obtaining information,
“Video/image format” such as Youtube, Instagram, and video advertisements,
“Audio format” such as radio, Audible, and audio distribution apps,
“Text format” such as blogs, Xwitter, and books,
And I often “ask AI” recently, but among these, I think “video/image format” is the mainstream worldwide.
90% of the time, you don't need it.
The fast movie mentioned above also applies to that, and the supply amount is outrageous as if the smartphone population = video/image supplier, and they are desperate for video production in the field of AI development.
Even that “video/image format”, in Japanese videos, there is an extremely large amount of “text transcription”, and it is virtually a text + audio format. I think that one of the reasons why the Xwitter population is large from the world’s perspective is that.
Conversely, it can be taken as a poor video expression, but after all, I think that the format of putting my thoughts into words is a relatively easy way to feel good. It’s fast ecstasy.
That's one of the reasons why I'm writing such a long story even though it's a preamble to a preface. And that reason is also told in the 7th volume of "March comes in like a lion (written by Chika Umino)" (around the 9th episode of the 2nd season in the anime) (abbreviated below
----------------------------------
Based on 1/100 of a pixel on the longest side
・Last time’s postscript - I digress again as soon as I say it
In a previous article, I speculated that the reason for the ban was due to secondary creations by AI, but I realized that there could be two more reasons.
That is the mosaic I often mention, and realistic illustrations.
Personally, I can't seem to get out of the uncanny valley.
The reason I first thought that secondary creation was one of the causes is because I referred to a person who is posting AI illustrations and is following me. That person’s illustrations had a lot of copyrighted characters, and I had the impression that they were applying more mosaics than necessary. So I thought the reason that person was banned was not because of the mosaic, but because of secondary creation.
However,
The impression I got from looking at many AI illustrators is that there are more people who are not interested in doujin activities until they come into contact with AI.
Of course, I think there are many people who, like me, were interested in doujin activities, but could not complete their works and did not reach the point of activity, but if I narrow it down to these two examples, I have the impression that there are more of the former people.
That’s because, at least in Japan, AI illustrations have become a “product” that says, “Let’s make a photo book with AI, let’s make pocket money with AI gravure and AI cosplay”.
I made this much with AI illustrations, you can make the same amount if you do as this product. There are things going around that are just moving to the part of making money with the product because they couldn’t make as much as they thought or couldn’t make it anymore.
These are the images that these people tend to create. Have you ever seen it?
There was a time when digital photo books using models close to live-action were overflowing on DL doujin platforms and Kindle (even the thumbnails and provocative words are similar), and AI YouTubers who use live-action-like images as standing pictures with the same reading software’s voice are also overflowing, and I think that’s the background.
Those people are interested in making money, but they are not originally from the doujin field, so they reach out to things that anyone familiar with doujin knows, such as “ero of Uma Musume is ban”, and above all, they do not understand the standards of mosaic.
Of course, in a previous article, I said that there is no clear standard for “mosaic”, but there are some guidelines and standards shown by each platform, and if you touch such works on a regular basis, you will notice that there is a difference in mosaic depending on the place of provision.
So, I think that those people who are not from the doujin field may have neglected the mosaic processing.
And as a characteristic of AI illustrations, things that are quite detailed in depiction and are subject to mosaic are expressed, so the judgment becomes stricter.
By the way, even in my past works, there is a mosaic processing that does not meet the correct standard due to a misunderstanding of the standard I recognized, so I think it’s not strange to be BAN at any time.
↓ From past images
And one more thing, in the past regulatory strengthening, even if it is not live-action, things with expressions close to it have also become subject to regulation, so I will explain later, but I think I will add two predictions that it may have been BAN from the characteristics of AI illustrations that tend to be realistic.
