“I have temporarily made all my posts on Pixiv private“It is a message that has been seen more and more on the Pixiv platform since last May 6. Illustrators with hundreds of thousands of followers on Twitter have been making this decision, motivated by a joint protest against the rise of AIs that generate illustrations.as well as the apparent support of pixiv towards these tools.
For example, “@Akmkmk3“, an illustrator with more than 240,000 followers on Twitter, posted an update on May 7 that he would stop posting new illustrations on Pixiv and would make all his previous posts private. In his words, “he will continue with this decision until Pixiv clearly expressed his position on the rise of AI.”
Many other illustrators, including “@Rswxx“with more than 640,000 followers, already”@kuroe16370547” with more than 130,000 followers, have announced similar measures. In both cases, the artists mentioned that they fear that Pixiv will continue to allow AIs to use their illustrations as learning models.
There are a lot of reactions to the now called “Illustrators Strike”, highlighting: “I’d like to support them with the goal of reducing the number of people bragging about illustrations they didn’t make themselves“; and “Now even the Internet is not a safe place, considering that an AI could use your published work.“.
The move to the protest of illustrators was caused by a rumor that emerged on social networks where it was claimed that Pixiv was allowing scrapinga technology that collects specific data from websites, in this case being used to collect all artwork published on Pixiv to make AI-powered artwork generation more efficient.
The technique discussed uses technology capable of creating new artwork from previously created artwork, which can result in images that resemble certain artwork by real artists. First, the AIs use a database of millions of illustrations. Next, through the instructions given by the user, the AI generates an illustration or a group of illustrations supported by the huge database at its disposal. This seems to have spread the impression on the part of the illustrators that Pixiv was not taking steps to avoid feeding the learning of the AIs, leading to their taking the current step.
However, in Pixiv’s terms of use it is explicitly prohibited to “the use of programs or tools that collect the publications made on the platform“. Simply put, scraping is, in fact, prohibited on Pixiv. However, it has been mentioned that the measures are currently inadequate and should be improved.
However, if the illustrations were also published on platforms other than Pixiv, this protest loses all meaning. Illustrators also seem to be aware of this, and some of them have made their illustrations private as an expression of their desire to ask Pixiv to strengthen its crackdown on AI learning.