Plum blossom "Ninety-Nine Chilling Picture" made with AI painting tools. Production: Li Shiqi, Meng Yang.
Plum blossom "Ninety-Nine Chilling Picture" made with AI painting tools. Production: Li Shiqi, Meng Yang.
New Year's pictures made with AI painting tools. Production: Xu Baojuan.
A dragon-shaped pattern made with an AI painting tool. Production: Xu Baojuan.
In recent years, with the emergence of human-computer interaction tools, we have ushered in the era of artificial intelligence (AI) technological innovation, especially the application of artificial intelligence in the field of art, which has brought about an artistic crisis under the background of artificial intelligence - how to establish the boundary of artistic creation? How to judge the originality of a work of art?
In my opinion, although AI painting has a major breakthrough at the technical level, it has its limits in terms of creation. AI is a technology and a tool from beginning to end, and AI paintings are not works of art.
The generation mechanism of AI painting works is a data-based algorithm, and the designer sets the model of its generation mechanism, and the AI painting tool finds out the regularity of a certain art from the algorithm and analysis of big data, generates the algorithm and interaction, and then further generates the image through the generalization, refinement, analysis and synthesis of the schema. From the perspective of artistic creation, this kind of artistic generation can only stay at the level of reproduction and imitation of painting, and cannot rise to the level of creation.
If you analyze it further, you will find that the AI painting tool is based on the imitation of the existing schemas, styles, and painting languages of art history, and further analyzes the styles and illustrations of these artworks through machine models, and then further imitates them through the method of neural networks, all of which have a premise: their generation principle is based on the imitation of existing artists' styles. In this process, AI painting can be copied, cut, modified, analyzed, collaged, appropriated, reproduced, interpreted, interacted, and feedback, all of which are just imitations, resulting in the direct result of AI painting's schematic copying, subject labeling, and homogenization of works.
It should be said that AI painting tools have become a huge database, stored in the cloud. From the perspective of the generation mechanism, the current AI technology can not only copy the old works of art, but also with the innovation and iteration of AI painting technology, AI painting programs can generate works of art based on the parameters of keywords. In particular, the AI painting model tool, which has become popular recently, is very simple to operate, and users only need to enter some keyword parameters to generate corresponding images through AI algorithms. In this generation process, the user can choose from different artists and art styles, such as Leonardo da Vinci, Van Gogh, Picasso, Dalí, etc., and can be generated in less than a minute. In addition, these tools can recognize lens and photographic terms, demonstrating their powerful ability to generate images quickly. Low threshold, easy operation, rapid generation, and superb imitation, all of which make this kind of AI painting tool "break the circle", as if entering the era of "everyone can be an artist", and then fall into the huge controversy of "whether AI painting is art". However, in essence, these AI painting programs are tools after all, lacking the artist's subjectivity and the originality of the artwork, and are just program iterations brought about by technological innovation. That said, AI painting tools are technology from start to finish.
The technological innovation of AI is tantamount to another technological innovation brought about by photographic technology. Just as photography is based on the reproducibility of imaging technology, AI is based on the infinite replication and reproduction of algorithms. Therefore, AI painting is only a reproduction and reproduction of existing artistic styles, lacking originality, and is only a medium for the integration of technology, technology and art, and has not yet risen to the level of art. AI has not yet generated the "subjectivity" of artists, has no independent perception, consciousness and emotion, and lacks real human life experience. All of this makes AI painting separate from real art.
Since its birth, art has carried the depth and emotional strength of human thought. The axis of artistic originality always revolves around the heart and spirit of the artist. No matter how AI technology innovates and how the times change, the essence of art is eternal, and what real art carries is the artist's unique life experience and the record of the artist's life emotion. The author still clearly remembers the heart tremor brought by Van Gogh's "Sunflowers", sunflowers are Van Gogh's body and soul, Van Gogh knows how to grasp the rhythm and rhythm of the brushstrokes, the restless and trembling brushstrokes are the burning of his passion for life, those sunflowers that seem to be still lifes are like human faces, expressing the theme of the struggle for survival, giving people a strong soul tremor. You can feel his despair, loneliness, pain, struggle, and love, and his golden sunflowers condense Van Gogh's scorching light of life.
In short, AI painting may cause a visual carnival, but it cannot bring a flutter of the heart. Nor is it possible for AI to replace human artists, because it cannot replace artists to think, experience, and empathize. Although we are in an era of technological acceleration, the iteration of technology is always changing, and its effectiveness is temporary and limited, and the key to the problem lies in how to adhere to the spiritual characteristics of art, the subjectivity of artists and the originality of artworks.
Author: Song Shilei is an assistant researcher at the Center for Aesthetics and Aesthetic Education, Peking University
Project team: reporters Zhang Yumei, Yu Yuanyuan, Xu Xinyi.