China News Service, Hangzhou, February 21st: Question: Zhejiang Cultural Observer: Will AI’s entry into the animation industry replace humans?

  Author Tong Xiaoyu

  Producing the approximately 4-minute "Journey to the West" animated short film originally took at least half a year, but now with artificial intelligence technology (AI), it only takes a week. Recently, the impact of AI on the animation industry continues to be fermented on the Internet. On February 21, experts said in an interview that when entering the animation industry, AI is both an assistant and an opponent. Many positions may be eliminated, but creativity will not.

Animation companies use AI for character design. Photo provided by Wu Jun

  On February 19, the AI-generated animated short film "Journey to the West" went viral on the Internet. It used AI to vividly present the story of "Journey to the West" from the birth of the stone monkey to the apprenticeship. From the Heavenly Court to the Flower and Fruit Mountain, the scenes shown in the video are meticulous.

  According to the producer of the animated short film, there are only four steps to use AI to generate a video: conceiving, generating static pictures with text, making the pictures move, and organizing and editing. While netizens are shocked by the exquisiteness of the video and the efficiency of AI, a question also arises. Will AI replace humans in the future when it enters the animation industry?

  "The development speed of AI is beyond my imagination. It is really making rapid progress." Ding Haixiang, dean and professor of the School of Animation and Digital Art of Zhejiang University of Communication, said frankly that from the current point of view, it is not even a replacement, but it will "liberate" a lot of people. Creativity brings a huge threat to traditional simple human creative work. "The lack of systematic originality is the fatal weakness of current AI. As a large-scale model tool, it requires prompts and organization. If the creator is not creative enough and only knows simple techniques, he will definitely be eliminated by AI."

  Wu Jun, general manager of Hangzhou Ruohong Culture Co., Ltd., has been engaged in animation work for nearly 10 years. His company has introduced AI into animation production in the past few years. "In fact, the entire industry uses AI, and we can only embrace it. Moreover, AI has obviously improved work efficiency."

  He gave an example. At first, the art setting of an animation took a team of more than 20 people more than 2 months to complete. Now with AI, a team of 5 people can solve it in 2 to 3 weeks, and the painting level is very good. , surpassing a painter with 10 years of painting experience. However, the flaws are also more obvious. The painting style is limited, and it is just a simple image design, which does not reach the level of a work.

The animation characters are compared successively through AI optimization. Photo provided by Wu Jun

  In fact, this flaw is being overcome. On February 16, the global star AI startup OpenAI released the Vincent video model Sora. Through simple text commands, AI can generate multi-shot long videos of up to 1 minute. The reporter saw that in the video generated by Sora, not only the scene was realistic, but the character's hair moved extremely smoothly and naturally.

  "This is a 'nuclear explosion' level work, and the shock level is far beyond imagination." Wu Jun said, believing that after Sora continues to integrate human works, its style will become more and more diverse, and what it outputs will no longer be a simple image. Rather, it is a complete work that realizes "one-stop" service of art, interpretation, script, and visual effects, just like a whole team producing it.

  The production is fast, the quality is high, and the production cost is low. It is a general trend for AI to enter the animation and film and television industries. As a major film and television province, since 2020, Zhejiang has ranked first in the country in terms of output and number of recommendations for cartoons. How does Zhejiang animation industry respond to the impact of AI? How to cultivate talents?

  Wu Jun believes that under the trap of algorithmic perception, although AI has intervened, creativity is still the top priority. "AI is a learning tool. What can't it learn? That's something that hasn't been created before."

  This also forces personnel changes in the animation, film and television industries. He said that in the future, comprehensive cross-professional talents will be more competitive in the industry. If you can only draw, have no creative ability, and do not understand the key points of the script, you may be eliminated.

  Ding Haixiang believes that the entry of AI does not mean that colleges and universities no longer need to cultivate animation talents. Intensive technical labor and simple repetitive production are the current advantages of AI. The selection of content direction, core creative concepts, key art styles and aesthetic choices of animation still rely on human thinking. In the past, colleges and universities paid more attention to the cultivation of students' professional skills and taught them how to put their imagination into practice. In the future, they will need to teach students broader literacy knowledge and aesthetic abilities.

  "It's like the same clothes, with different combinations for different aesthetics. People with strong matching skills and strong aesthetic abilities will have better results." He said that currently the school has introduced AI into teaching, but in the future training process of undergraduate talents On the one hand, we must guide students to actively use new technologies, on the other hand, we must control students' mindless and speculative use of AI, and teach students "lower limit" knowledge, that is, aesthetic ability and ways and methods to use tools to strengthen their own creativity. This is both AI The current shortcomings are also traps that students currently fall into when exposed to AI.

  Currently, Sora, the most advanced text generation video tool, already has the ability to think linearly and build simulated worlds, and its performance far exceeds that of early text generation video tools. Ding Haixiang said that this is no longer a simple text generation video large model tool, but a prerequisite for the arrival of the metaverse. As relevant professional educators in colleges and universities, we must actively embrace and pay attention to the application and progress of the latest technologies in the future. At the same time, professional learning in colleges and universities should be closely integrated into the development process of new technologies, with the aim of cultivating talents who can seamlessly connect with new technologies immediately, and truly cultivate compound talents with the characteristics of the AI ​​era when changes come. (over)