When AI meets art, how can AI better benefit creation? Experts interpret

AI technology is integrating into our lives at an unprecedented speed, reshaping the ways art is presented and communicated. As an auxiliary tool for artistic creation, AI is demonstrating increasing vitality and creativity in multiple fields. Art itself is constantly injecting meaning and soul into AI. What kind of spark will be ignited when the two meet and collide?

This is the first AI symphony in China, ‘Thousand Miles of Rivers and Mountains’. The combination of ancient famous scrolls with musical poems complements each other, making this enduring masterpiece even more captivating. It is hard to imagine that this symphony is entirely generated by the self-developed artificial intelligence music model of the Central Conservatory of Music.

These clips are all sung by AI, sounding almost no different from human singing. In the fields of lyrics writing, music performance, orchestral composition, etc., AI is becoming a new creative force. Style, instrumentation, harmony, and lyrics can be freely chosen, and a complete song can be written in 23 seconds; high-precision music source separation technology enables free switching between vocals and accompaniment; moreover, AI robots can also sing, play, and conduct.

Even in the metaverse, the audience can embody AI digital beings, stand next to the singer, and enjoy an immersive concert.

Director of the Department of Music Artificial Intelligence, Central Conservatory of Music and Director of CCF Computing Art Branch, Li Xiaobing: The emergence of artificial intelligence has brought about significant changes in the six secondary disciplines and majors of music, such as the polyphonic generation model, which may be generated in a few seconds.

Exhibition: Let the work “live” and “move”.

In the nearly 10,000 square meters of exhibition space at the Beijing Wu Xiang Art Space, when you take a photo of ‘flowers’, the flowers will scatter. When you take a photo of ‘light’, colorful neon lights will scatter. Children draw colorful animal drawings with colored pens, and the artificial intelligence system in the venue learns and scans in real-time, simulating the visual processing methods of human eyes and brains, and generates small animals with significant features corresponding to the original drawings. They seem to have been given life, appearing on the walls, trees, and ground, playing hide and seek and chasing with the children.

Yang Wenting, Director of Beijing Wu Xiang Art Space: Frogs eat butterflies, lizards eat frogs, snakes eat lizards, salamanders eat snakes, and thus a food chain is formed. This allows children to create, learn while having fun, and experience scientific principles.

Video generation: applied to fields such as film and animation production

The panda wearing sunglasses plays rock music below the snowy mountain, while the girl holds a goldfish in her hand, and the fish swims in her hand. There is no need for a camera to shoot, no need for post-production, just input the command text or upload an image in the AI video generation model to get vivid video content. A five-second video can be produced in one or two minutes. Currently, in the fields of film and television production, video animation, and many other areas, China has multiple large AI video generation models.

Online Literature: AI-Assisted Creation of the First Million-Word Novel

In the field of literature, AI has also provided a brand new writing mode. Recently, East China Normal University used AI to create the first million-word artificial intelligence novel in China in the form of ‘artificial intelligence large language model + prompt word engineering + manual post-editing.’ The human designer designs the story framework and prompt words, while AI completes the character relationship map, batch paragraph writing, and perfecting story details. A diligent writer would need a year and a half to complete such a work without a break, while using AI only requires one and a half months.

Archaeology: Breaking through traditional techniques to build ‘models’ to decipher oracle bones

In the field of archaeology, AI also provides researchers with more decoding methods. Through platforms such as Oracle Bone AI Collaborative Research, three-dimensional modeling, micro-trace enhancement, artificial intelligence, and other technologies based on the “Oracle Bone Full Information Model,” breakthrough the traditional reliance on rubbings, copies, and photos in research. It not only allows for a clear view of oracle bone inscriptions but also reveals stroke overlays and modifications. By accurately analyzing stroke directions, character outlines, and matching characters, valuable clues are provided to unlock the ancient civilization code from over 3000 years ago.

Is AI+art students’ content protected?

AI creations bring convenience and innovative vitality to the cultural field, but also bring some new issues. For example, does AI-generated content have copyright? Should the corpora used by AI for learning be defined as fair use, or require copyright authorization? Will the data itself affect security?

Recently, the first instance of the “AI-generated image infringement” copyright dispute case in Wuhan was pronounced. In May last year, Mr. Wang used AI software to generate an image and posted it on a social platform. However, the following month, Mr. Wang found that the image appeared in an advertisement for an AI painting training camp of a certain technology company. Mr. Wang believed that the company infringed on his rights and sued it. In the end, the court found that the sued image had original expression and should be protected by copyright law. The first instance ruled that a technology company in Wuhan should compensate Mr. Wang for his economic losses and reasonable expenses of 4000 yuan.

Judge Xudongdong, a second-level judge of the People’s Court of Donghu Science City, Wuhan East Lake New Technology Development Zone, stated that first of all, it should be examined whether the keywords and parameters input by the plaintiff Wang meet the standard of originality; secondly, the relevance between the plaintiff’s creative activities and the creation of works. The plaintiff inputted dozens of keywords when creating, such as scrolls, light and shadow effects, distant views, etc., underwent dozens of outputs and selections, and finally obtained the relevant images.

Does AI-generated product have copyright?

Currently, from a legal perspective, there is no clear provision for the protection of AI-created works.

Professor Wang Feng, Dean of the School of Communication at East China Normal University: In the end, does copyright belong to the user or the platform? We temporarily cannot obtain legal protection, which is also a big issue at present.

The Executive Vice President and Secretary-General of the China Written Works Copyright Society, Zhang Hongbo: We have more than 2,000 radio and television stations, more than 580 publishing houses, and more than 400,000 books published each year in China. There are also more than 10,000 newspapers and magazines. Many of these works are protected by copyright. If these works are used by large, technologically advanced artificial intelligence development companies for model training, we should pay attention to the authorization of these copyrighted works needed for artificial intelligence corpora, and not deny the legitimate rights and interests of other content producers. We should pay attention to the interests of various nodes and groups in the industrial chain, so that the law can adapt to the development of the industry and the industry can comply with the legal provisions.

How to Break the Dilemma of “AI Illusion”

In addition, the content processing ability of AI, especially the so-called AI illusion dilemma based on complex information sources, has also attracted much attention.

Wang Feng, Professor and Dean of the School of Communication at East China Normal University: The difficulty of creative writing lies in maintaining the consistency of the logical structure of the article. The ability of large language models currently does not reach the level of human writers.

Shanghai Jiao Tong University’s Artificial Intelligence Governance and Legal Research Center Secretary-General, Director of China Internet and Information Law Society, He Yuan: AI cannot achieve complete authenticity now. For example: voice and image forgery, putting one person’s head on another person’s body. There is also algorithmic discrimination, where algorithms themselves have a certain bias, especially in the cultural field. How to ensure high-quality datasets that conform to our socialist core values, and feeding them to large models, is also very important.

AI+Art Tool Empowers Service Creation

The 55th “Statistical Report on the Development of China’s Internet Network” released by the China Internet Network Information Center shows that as of December 2024, the user base of generative artificial intelligence products in China has reached 249 million, accounting for 17.7% of the total population. There is no doubt that AI has entered people’s lives on a large scale, but how should we define the relationship between AI and creators, and how should we interact with AI?

Zhang Hongbo, Executive Vice President and Secretary-General of the China Written Works Copyright Society: The most basic boundary of artificial intelligence should be to make technology benevolent, to serve humanity, and to bring great, disruptive, and transformative convenience to our work, study, and life, in order to stimulate the enthusiasm of a large number of literary and artistic workers and content creators for creation.

Shanghai Jiaotong University Secretary-General of the Artificial Intelligence Governance and Legal Research Center, Director of the China Internet and Information Law Research Association He Yuan: AI is essentially a tool, similar in essence to traditional computers and the Internet. It can promote people to create better, emphasizing human decision-making power, control, and innovation.

Li Xiaobing, Director of the Department of Music Artificial Intelligence at the Central Conservatory of Music and Director of the CCF Computational Art Subcommittee: You are a connoisseur, a selector, a guide. Therefore, you may be able to create more exquisite art. The emotional exchange between humans cannot be replaced, but machines can help intensify my feelings. In a sense, artificial intelligence has empowered various industries, so we must embrace it.

(Source: CCTV News)

Source: East Money

Author: CCTV News