With the continuous development and application of artificial intelligence technology, AI painting, as an emerging form of artistic expression, is constantly impacting the traditional art field.

The high auction price of the AI painting work "Portrait of Edmund Bellamy," "Unfinished - To Be Continued," and "Space Opera House" winning the top prize at the U.S.

Art Expo of Colorado, along with the introduction and popularity of AI painting tools like Wenxin-Yigue, Midjourney, Novel AI, Disco Diffusion, Imagen, among others, are all announcing the arrival of the era of AI painting art.

AI painting undoubtedly lowers the threshold of art creation and improves the productivity and inclusiveness of art. Nowadays, AI painting is no longer a niche medium in the field of generative art, and the open-source nature of the program makes it accessible to everyone. "Text-to-image" (text-to-image) allows access to AI paintings by entering descriptive text.

Although the current AI technology is not yet able to perfectly reproduce the descriptive text entered by the user and faces some technical limitations, such as not being able to deal well with certain complex scenes, detailed movements, or subtle color changes due to algorithmic shortcomings, people are still immersed in the fun of participating in generating art.

During its advancement and use, AI technology has gradually replaced many heavy, mechanical, and low-skill jobs. The art of painting was once considered an area that was not easily encroached upon by AI and was relatively safe due to its unique expertise and creativity.

However, the large number of AI painting works emerging and gaining popularity has led many art creators and scholars to worry that people, in the process of using AI, will be reverse domesticated into the servo mechanism of technology, leading to the degradation of their own creative abilities. The theory of the demise of the arts is once again in the air.

The essence of AI painting lies in machine learning, which requires training by borrowing image datasets of existing artworks, which can be crawled over the Internet without the permission of the original creators.

While AI painting does not directly copy existing works, it draws "inspiration" from the artwork that trained it, and the original creators cannot reclaim their work or prevent the AI from acquiring more. This also draws attention to the ethical crisis of AI paintings in the field of copyright.

The copyright issue of AI paintings can be discussed in terms of attribution and infringement. In contrast to traditional paintings, created by human artists with clear attribution of copyright and intellectual property rights, the attribution of AI paintings is difficult to ascertain.

This difficulty arises because AI technology models, programmers, artists, and end-users may all have an impact on the process of creating the work. Assuming that a work generated by AI is copyrighted, who can own it?

The Copyright Law first excludes the possibility of AI owning the copyright itself, specifying that copyright holders only include "authors, other citizens, legal persons, or other organizations enjoying copyright in accordance with this Law."

Based on this, some scholars argue that programmers and end-users are candidates for copyright in AI-generated works.

The utilitarians among them argue that the purpose of copyright law is to incentivize authors to create expressive works for the public good, and that granting copyrights to end-users is most likely to achieve this incentive; or simply to make copyrights in AI-generated works unencumbered, so that these works can be freely shared and used by the public.

This view has inevitably met with strong resistance from art creators, who insist that once their work has been used as a training dataset for AI paintings, the copyright of their generated works should not bypass them.