AI-generated Journey to the West Short Film Stuns Netizens, Sora Poised to Revolutionize AI Video Industry
-
A year ago, the emergence of the AI model ChatGPT sparked widespread discussion and attention across society. Time flies, and the AI industry has undergone earth-shaking changes. Recently, a blogger from Sichuan named "AI Madhouse" released an AI-generated animated short film based on Journey to the West online. Its stunning visual effects left many netizens in awe, marveling at the rapid advancement of technology.
AI Handles Most of the Process, Even Storyboarding Relies on ChatGPT
This 3-minute and 56-second animated short film is based on the first episode of the original Journey to the West. Using AI technology, it vividly presents the story of the Stone Monkey's birth and his journey to seek apprenticeship and learn skills.
The celestial court in the video is highly recognizable, with palaces built along the mountains, arranged in an orderly yet varied manner. The roofs are covered with golden glazed tiles that shimmer in the sunlight, instantly conveying a mysterious and solemn atmosphere. The Flower-Fruit Mountain, on the other hand, is filled with tranquility and harmony. Lush trees, cascading waterfalls, and quaint stone bridges together sketch a picturesque utopia. The video also showcases scenes such as Sun Wukong sailing across the sea, arriving in a town, and heading to the Lingtai Fangcun Mountain, leaving netizens amazed. "This video is truly breathtaking! The AI-rendered Journey to the West is absolutely incredible! Just the opening scene alone is stunning!" "The scenes are indeed spectacular. I hope the creator continues to update along the storyline. Thumbs up!" In addition to 'Journey to the West,' Mr. Feng has used AI to create video animations such as the Silk Road, the Nine Sons of the Dragon, and Cthulhu Mythos. Currently, he has over 100,000 followers on Douyin and has received more than a million likes. In the comments section, while many are amazed, some have asked how these videos are made. Mr. Feng told reporters that generating videos with AI involves four steps: conceptualization, generating static images from text, animating the images, and editing.
"The first step is definitely conceptualization. Taking the first episode of 'Journey to the West' as an example, from the chaos of creation to the birth of the Stone Monkey and then to learning skills from a master, the first thing is to determine how many scenes and shots are needed." Conceptualization may seem to require 'wild ideas,' but in reality, it doesn't. Mr. Feng explained that he uses ChatGPT to analyze the original text, letting the AI help plan some shot ideas, from which he then selects the suitable ones. After planning the shots, he uses AI drawing software to create the desired images and then employs 'image-to-video' software to animate them. Both processes typically generate a lot of discarded drafts. "Currently, AI has difficulty fully understanding my intentions. Whether it's images or animated clips, the randomness is quite high. Each time, I generate a lot with AI and then pick the suitable ones from these materials to proceed to the next step."
"The remaining step is to put the materials together for editing. I come up with the dialogue myself and then use AI for voiceovers. For some special characters, I do the voice acting myself, using a voice changer, and can further refine it with sound effects during editing." Adding transitions and background music, an AI-generated animated video is born.
Some netizens want to see fight scenes, but currently, AI-generated videos still have many limitations. "Two years ago, I used an AI drawing model released by Google. At that time, AI-generated images were extremely abstract and fragmented. For example, when drawing a person, the AI could depict facial features, but they would be severely distorted. Everyone criticized it, saying the software could only produce horror images. However, by the end of last year, the model had been updated to its fourth version, showing significant improvement. I was even invited to participate in the beta testing," Mr. Feng explained. Now, the AI drawing model has advanced to creating various styles such as cartoons, sketches, and watercolors, with even greater attention to detail. "For instance, it can now depict embroidery on clothes, even intricate gold threads."
"Since its inception, AI painting has sparked numerous controversies. Illustrators argue that AI uses their works for training and is now competing with them for jobs," Mr. Feng admitted. The rapid iteration of AI technology has indeed affected some people's livelihoods. "But there's no stopping it. It's like Pandora's box—once opened, it can't be closed again. Resisting it is pointless."
"Take the animated short film 'Journey to the West' I recently released as an example. I generated over 3,000 images using AI and ultimately selected about 100. In the past, even a seasoned artist would need at least six months to produce these drawings, let alone animate them—that would have been unimaginable."
In the comments section, many netizens expressed interest in seeing battle scenes from 'Journey to the West.' "Currently, AI-generated videos still have many limitations, and the technology isn't there yet," Mr. Feng said. Many things that were once impossible are now achievable with AI, and what seems impossible today might become possible in the future. "When AI becomes powerful enough, I will definitely fulfill the requests of my audience. I hope that day comes soon." Extended Reading: Sora May Revolutionize the AI Video Field
Reporters noted that OpenAI recently released the video generation model Sora. According to the introduction, Sora can generate up to one-minute videos based on user prompts, text instructions, or static images, maintaining high visual quality. These videos include intricate scenes, vivid character expressions, and complex camera movements. Additionally, Sora can extend existing videos or fill in missing frames.
Currently, dozens of demo samples are available on Sora's official website. These samples not only accurately present details but also understand the physical existence of objects in the real world and generate characters with rich emotions.
Many industry experts are optimistic about Sora. Zhou Hongyi, founder of 360 Group, stated that the birth of Sora means the realization of AGI (Artificial General Intelligence) could be shortened from ten years to just one or two years. "Once AI is connected to cameras and watches all the videos online, its understanding of the world will far surpass that of text learning... What it demonstrates is not just video production capability but the success and breakthroughs that come when large models understand and simulate the real world."
Liu Xingliang, a member of the Expert Committee on Information and Communication Economy at the Ministry of Industry and Information Technology and a digital economy expert, said this marks a new era for AI technology in content creation. "Sora can help creators turn their ideas into reality at lower costs and faster speeds while providing audiences with richer and more diverse visual experiences. This significant leap in technological innovation indicates that AI will play an increasingly important role in all aspects of human life in the future." Mr. Feng is also very optimistic about Sora. He stated, "Although I haven't used this program yet, judging from the official descriptions and the videos released, the elements generated by this model show a high level of coordination. Moreover, it far surpasses current models in areas such as storyboarding, motion capture, and transitions between close-up and wide-angle shots. In other words, the challenges we face today may no longer be issues in the future."