AI riding face, artist protest

Author:Institute of Physics of the Ch Time:2022.09.11

Since last year, "AI painting" is frequently attracted people's attention. It seems that for a short period of time, AI creations from different projects will suddenly appear and swipe the screen on social media. With amazing colors and artistic styles, a wave of "wisdom crisis is uncertain". Back to your own field secretly developing.

Sometimes it may even cause the problem of public opinion with false chaos

But now many people have been unable to stay as simple as their existence -a few months ago, some people lamented that the artist was going to face the unemployment crisis. Now this crisis seems to have arrived.

In August, at a large -scale art expo held in Colorado, the United States, the organizer awarded one of the digital art awards to a set of works that looked quite magnificent.

The work is called "Space Opera House". Friends who are familiar with many recent paintings of AI are not difficult to see the typical AI painting traces. The subsequent explanation of the winners also confirmed the origin of the work.

According to the author's own statement, the original manuscript was produced by the recently hot painting AI MIDJOURNEY. He rely on keywords to generate 100 pictures. After several weeks of fine -tuning and screening Printed and sent it directly to the judges of the event.

The one shown above is one of his award -winning works.

After this incident was exposed, it caused controversy from netizens and professionals in a short time. The reason is not difficult to imagine: the author himself defeated other creators who spent time and energy, and got the judges from the judges. Recognize, can this really be an artist?

"Someone participated in the game with AI, and took the first place. This is too bad."

However, the author also expressed dissatisfaction with public opinion, and it was right to firm himself. He claimed that not only did he have himself in the signature of his work, but he also clearly wrote painting AI Midjourney. He also reminded the official judges that the main body of this series of paintings was completed in AI, but the other party still decided to award the award to him.

Both parties hold their own words. Netizens criticized the behavior of low cost and disrupted rules with AI painting, thinking that this was "tentative bottom line" and "driving bad heads"; It means "the art world must admit AI art after all."

In fact, this event is not a big prize, and the winner has only won only $ 300, but the problem of use of AI works caused by it quickly caused the media to report. Among everyone, the tension of professional artists cannot be ignored.

A month ago, when the American "Atlantic Monthly" reporter Charlie Woz was not used to match his article, he did not choose pictures from the picture copyright website Getty Images as usual, but used AI MIDJOURNEY to paint. In the end These contents were published together.

Pictures used in articles

As a result, this article was a manuscript that criticized the extremely right -wing anti -LGBT+person. The most notable content became these pictures that marked "from Midjourney".

The second day of the article was published, and many well -known artists in the industry collectively protested against the Atlantic Monthly. They expressed shock and depression, thinking that a national newspapers and magazines took the lead in using AI instead of paying to the painter to obtain pictures. It is difficult to protect other peers, and the result may be that the original customers have used this excuse to reduce the art budget.

This statement looks very impactful, so a humble thing has quickly triggered the word of people. After being strongly attacked, Charlie Woz himself quickly issued a mistake, claiming that "never do this again."

In order to show sincerity, Charlie Woz deliberately asked the well -known American cartoonist Matt Bos to understand the possible impact, and the other party explained this:

"More and more technology is created to do basic work and save costs for the rich, but most of them have not benefited the public ... AI technology is one of them. It can help some developers help some developers And technicians, but for the illustrator, this is a frustrating development. "

Is this words worried or imminent? The user's inadvertent move can cause the worries of the creators. Such a bad response is the far -reaching impact of AI painting.

Human artists' distrust of AI painting is destined to be difficult to escape. In addition to "grabbing rice bowls", the painting AI that grows up by "eating a hundred family rice" also has many serious issues that involve plagiarism and misappropriation.

Just last week, a illustration named MIMIC appeared in Japan to generate AI, which once caused collective protests from many artists.

MIMIC introduced itself in its official account: "It allows artificial intelligence to learn the artist's personality style and create an exclusive illustration maker that can create this style"

As long as the AI is provided with more than 30 plans, netizens can train a exclusive painting AI. When describing their target users, MIMIC also said that "it helps to enhance the painter's understanding of his painting style."

It sounds good, but many people see the word "plagiarism" from the lines -because MIMIC claims that AI's illustrations generated by identifying these paintings will have the official watermark of MIMIC. In other words, the copyright of these generated maps is in their hands. Soon, a crowd of painters believed that MIMIC's purpose was impure, thinking that their own work style may be officially obtained, and turned to misappropriate NFT technology or other business fields to make profits.

As a result, everyone began to publicly oppose MIMIC, and the resistance of resistance quickly spread in the circle. This wind direction even affected the entire field of AI painting. The artists began to call for everyone to put watermarks in a prominent position, and even very negatively deleted all their works on social media to prevent them from becoming a AI learning "material" to show resistance.

It is worth mentioning that some netizens have proposed an interesting "countermeasure" method, which is to "feed" a large number of Disney or Nintendo's character pictures to MIMIC, let AI generate works with watermarks, and solve this artist to solve this painter. The problem that can't be thrown to the top legal department of the Western Hemisphere and the Eastern Hemisphere.

There are many emotional situations of people's reactions, but it also reveals that AI painting may exist, including plagiarism, including plagiarism.

According to the Japanese Copyright Law, the artists are currently not prohibited from being learned by AI in data mining. They can only ask the works to not be uploaded to some specific artificial intelligence websites.

This is also the reason for everyone to oppose MIMIC so strongly: In a sense, existing copyright laws will not have any restrictions on the learning environment of painting AI, but if the content of the artists' own work is referenced and used by the AI, But there is no specific statement or copyright protection. The only way to resist the artists is to keep their works away from AI as much as possible.

In fact, this is not only the problems existing in Japan. The AI painting instead of artificial or even profit. Many troubles are destined to be in front of the vulnerable painter group.

The contrast between real artists and painting AI is indeed a bit cruel in terms of work efficiency.

A painter who takes ten years to grow up, from composition to completing a business illustration, often may take more than ten hours or more. Today, the well -known Midjourney, from entering keywords to getting paintings, may not even need for a minute.

As mentioned earlier, there is no need for too many artistic skills. A work that is obtained by only a few keywords. As long as you make some adjustments, you can also compete with the high -level works completed by the real person. Essence

And there are many AI painting tools like Midjourney -such as Stable Diffusion. With the support of the artificial intelligence team behind it, its painting can only be described as "falsely false". Especially in terms of character details, if the description of the input is accurate enough, it can indeed reach a glance at a glance.

Like a "reverse Turing test" for human beings

Unsurprisingly, Stable Diffusion, a powerful painting AI, also inspired many painters to grab the crisis of rice bowl and copyright. They even reported the title of an unofficial Twitter account of this AI.

But of course, the artificial intelligence that has already "ran up" will certainly not stay for whom. In China, there are also recently popular domestic painting AI texts, which can be used as a typical two -dimensional painting style image according to text output. Although the details lack a lot, the color and style of painting are not lost to many amateur artists -this is only the result of a few seconds.

More importantly, AI is still learning and "growing". The update speed is fast, and no one knows how the content they produce will evolve in the future. It is not difficult to see that although the AI painting at this stage is still unrealistic, the worries of some artists are obviously unreasonable.

When these paintings are quite abstract and artistic, and the paintings may be rough and fine but without flavor, they have been directly used to use it in front of the stage. Even when the competition stage and artificial painting are higher, the uneasy emotions are intensified. It became a reasonable thing.

As for the legal issues that may exist in the commercial field in AI painting, it is even more like an indescribable black hole. The artificial intelligence team behind each painting AI is very ambiguous in the copyright and legal issues of the output works. Whether it is the creator of the AI reference or the consumer who accidentally purchased an AI painting, the two parties are currently It is difficult to find specific objects to protect their rights.

Moreover, the application value of AI painting is no longer avoided. In many fields (such as game art), AI can already exist as auxiliary tools. AI like Stable Diffusion can provide accurate scene perspectives or lighting treatment. Many painting processes that can be done in the past can also help the creator to help.

Just as the changes in the modern industrial system have replaced manual technology in the past century, the same story has also occurred in the artist group in the era of artificial intelligence, which is similar to the birth of the camera that year.It is just that no one knows whether someone really is abandoned by technology and the times this time.Reprinted content only represents the author's point of view