Recently, the discussion on "Artificial Intelligence Internet Literature Creation" initiated by this edition has continued to attract attention, and many scholars have sent written articles on this topic to explain their respective views and opinions. This issue has two articles selected for publication. Among them, Professor Ouyang Youquan of Central South University believes that no matter what stage artificial intelligence evolves to, there are inevitable artistic limitations. At the same time, ethical boundaries should be set for artificial intelligence writing, and it cannot be allowed to develop into a tool that "backfires on humanity." Young scholar Ma Xiaoyan pointed out that the content of artificial intelligence writing is highly mechanized and homogeneous, and it is difficult to generate vivid realistic concerns, and it is impossible to promote the evolution of a "new literature" with real emotions, dialectical thinking, and ethical significance in a practical sense.

The application of artificial intelligence in the field of literary and artistic creation is making writing poetry, novels, composing music, painting, video production, etc. easy and comfortable. The threshold for artistic creation has been lowered and is no longer exclusive to humans. Will this promote the progress of art, or cause aesthetics to go off track? Is the "art" produced by the "technical Rubik's Cube" still the spiritual home that allows us to "dwell poetically"? The artistic ethics issues among them urgently await our thinking and answers.

1. Artificial intelligence makes artistic creation simple and easy in the past.

The artificial intelligence born in the 1950s was not born for literature and art. It is used in literary and artistic creation, but it is a functional presentation of "artistic technology". Judging from the actual effects, the impact of artificial intelligence on literary and artistic creation is undoubtedly huge and far-reaching. If the poetry creation software that appeared in the early years, such as "Orion Star", "Daoxiang Laonong Poetry Machine", "Microsoft Xiaoice", "Yuefu" and "Jiuge", in addition to making people feel novel, are not enough to "steal" poets' jobs, then the situation is very different after generative artificial intelligence (AIGC) enters the field of artistic creation. It has the potential of "the future is here, and art is changing." Whether it is literary writing or artistic production, they all show amazing creative potential and mass production levels.

New creations such as AI novels, AI music, AI paintings, AI movies, and AI short videos are rapidly emerging. Artificial intelligence not only puts the right to speak in art into the hands of the public, but also simplifies the complex creative process into text prompts, machine generation, word-based texts, text-based pictures, text-based videos, and "text picture" generated videos. As long as the demand information is given, the intelligent program will present what people want before their eyes like an "Aladdin's magic lamp." This turns the highly specialized literary and artistic creation of the past into an easy game that everyone can participate in.

Not long ago, "Tokyo Streets After the Rain" created by Vincent Video Model shocked the world; using AI music generation tools to create music, a person who does not understand music can also match the melody to the recipe and turn it into a catchy song. There is also a 90-minute film "Terminator 2: Judgment Day" entirely completed by artificial intelligence. It premiered in Los Angeles, my country's first A The full-process micro-short drama "Chinese Mythology" was launched on CCTV. The "Large Model Novel Creation System" used for literary creation also began online testing in March this year, and a million-word novel "The Apostle of Destiny" was created... Many friends around me are trying to use artificial intelligence to create poems, portraits, songs, or generate short video products. The suddenly lowered threshold for creation has made artistic creation that was previously out of reach simple and easy, and works within reach.

The artificial intelligence sketching device draws sketches through on-site shooting. Xinhua News Agency

2. "Emotional emptiness" and "absence of meaning" will become huge challenges for AI creation to confirm its own artistry.

The artistic trend of the layout of artificial intelligence creation artifacts is calling for the arrival of a new era of art. When an intelligent tool develops from a "technical means" to a "creative companion" and then to an "art master", on the one hand, we are delighted by the creative potential of cutting-edge technology, but at the same time, we also need to maintain a clear understanding of the artistic worries brought by artificial intelligence.

For example, AIGC, as a "quasi-subject" in artistic creation, cannot escape the inherent shortcomings of "silicon-based life": lack of consciousness and self-awareness, and no life experience that is indispensable for artistic creation. The former makes it impossible to acquire intellectual cognition and rational judgment about the world, lack the consciousness, initiative and moral self-discipline to pursue goals, cannot understand and absorb the historical experience accumulated by human beings from an independent value standpoint, and cannot independently form a subjective relationship between itself, society, others, and specific contexts. The aesthetic relationship between bodies and objects can only be combined into sentences or paragraphs based on the training model of algorithmic probability and computing speed, or the arrangement and combination of graphic (image) information using bit pixels based on certain instructions, without understanding the meaning of these combinations and why such answers are provided.

"Silicon-based life" makes AI creation "lost its roots". The pseudo-subject does not have the joy, anger, sorrow, love and hatred derived from the practice of living bodies. There is no "iron threshold" as Wang Euzhi said, "what you experience with your body, what you see with your eyes". The historical mirror, the spirit of the times, and life experience that are indispensable for artistic creation are impossible to talk about. The data used for large AI models is nothing more than "second-hand life" rather than "physical experience", and the "informed intention" it expresses comes from "feeding by others" rather than one's own "life adventure". This subverts the creative principles of "devoting oneself to writing" and "art to clarify one's ambition" in literary and artistic creation. It cuts off the organic connection between the work and the artist in relation to society, history and life, and is bound to rob the meaning of the existence of art. Therefore, the lack of life experience and physical perception are fatal flaws in AI creation. The "emotional void" and "absence of meaning" it brings will become a huge challenge for AI creation to confirm its own artistry.

Based on the above two reasons, another more important limitation of AI creation will be the lack of artistic originality. The answers provided by generative artificial intelligence are matched from existing corpus resources based on algorithmic logic, and are the result of clever selection by "borrowing" intellectual property resources from previous generations. In other words, nothing it obtains is original. It is both limited by “existing” and missing from “unknown”. Although artificial intelligence is "smarter" than humans in many individual skills, computing programs without human emotions and autonomous consciousness are ultimately unable to form original capabilities. It can be "fed" based on human information and pre-trained with big data to complete the expansion from "1" to "99", but it is difficult to achieve a breakthrough from "0" to "1", which is more obvious in artistic creation.

For example, relying solely on technical intelligence to produce AI Shakespeare and AI Kafka is nothing but a pipe dream. Art comes from life, and AI art comes from corpora; artists create with their souls, and AI expresses with programs; the "roots" of art come from life and human experience, while AI creation is a cold "painless operation." Because of this, without the guidance of historical sages, artificial intelligence cannot understand the symbol of Qu Yuan's "vanilla beauty" and cannot discern the meaning of "a good song" in "A Dream of Red Mansions". If an intelligent robot were to replace Armstrong and land on the moon for the first time, it would not have any heartfelt joy and pride. In this sense, this is why "Chang'e's moon is more attractive than Armstrong's moon."

Artificial intelligence robotic arm performs text creation and performance. Xinhua News Agency

3. Artificial intelligence must fulfill the artistic mission and ethical responsibilities of human subjects

It is worth noting that the artistic concerns of artificial intelligence creation are based on the functional judgment of the current level of artificial intelligence, but it is not the ultimate ethical presumption of AI artistic creation. In fact, artificial intelligence is divided into levels, and the functional performance of different levels will be very different. Researchers generally divide generative artificial intelligence into three levels from low to high, namely weak artificial intelligence, strong artificial intelligence and super artificial intelligence.

Among them, weak artificial intelligence is also called "algorithmic intelligence", which refers to program intelligence that can only complete specific tasks according to algorithmic logic in specific fields. It is called "single superman". For example, defeating the world's top players in the field of Go, intelligent navigation can accurately position in complex spaces, and autonomous driving technology can detect environmental information in real time and provide the best solution in a timely and accurate manner. Strong artificial intelligence is a kind of "linguistic intelligence" that can perform speech recognition and semantic understanding according to the user's instructions. It can also think, make decisions and learn like humans. Its intelligence level can be the same as that of humans, so it is also called general artificial intelligence, or "Homo sapiens-like". At this time, AI has possessed human wisdom and self-awareness. It can not only understand and analyze complex problems and propose solutions, but also continuously optimize and improve its performance in compliance with human emotions, ethics and moral concepts. Generative artificial intelligence is the product of the transition stage from weak artificial intelligence to strong artificial intelligence.

Super artificial intelligence is called "imaginary intelligence", which refers to intelligence that can far exceed human beings in all fields, and its intelligence level far exceeds human thinking ability. Because super artificial intelligence occupies a high dimension and can eliminate the meaning of human existence, it is extremely complex and potentially dangerous. Its operation may produce results that humans cannot understand or control.

AI art is created by intelligent "quasi-subjects". Does this "quasi-subject" want to fulfill the artistic mission and ethical responsibility of human subjects? The answer is yes. This is true for the creation of weak artificial intelligence, and it should also be the same when it comes to the creation stage of strong artificial intelligence or even super artificial intelligence. The reason is that weak artificial intelligence is used for auxiliary creation of "human-machine" interaction. Human dominance controls intelligent programs like arms and fingers. Human moral beliefs and ethical norms, human values ​​​​cognition and judgments of truth, goodness and beauty not only determine the design of big data intelligent models, but also restrict the preposition prompts of "quasi-subject" creation.

Even if it develops to the stage of strong artificial intelligence or super artificial intelligence, and "quasi-subjects" become "art masters", then AI art may show a completely different look to people. Will it definitely ignore or trample all the "logos origins" given to art by humans? The power of human reason and its artistic beliefs gives us reason to believe that any intelligent system developed in the future will ultimately serve human will and purpose and be a technical extension of human intelligence. If a certain kind of "smart machine" upgrades into a tool that backfires on mankind, it must be a human problem, not a machine problem. Humans should use rational laws to develop technological "gates" or formulate legal regulations to anchor the development direction of this technology, so that it can serve human civilization and social progress, and not allow it to go astray and harm humanity itself.

Based on this, if we admit that artificial intelligence art is the technical externalization of human artistic creativity, then there are reasons to set ethical boundaries for artificial intelligence creation, such as the identification of truth and falsehood, good and evil, beauty and ugliness, recognition of moral values ​​such as compassion, awe, and sympathy, and the pursuit of lofty ideals and poetic realms.

"Guangming Daily" (Page 9, July 20, 2024)