Image and video startups are starting to use AI to make money. How does AI reshape the work of practitioners?
Author：Eve Cole
Update Time：2024-11-23 16:00:01
From large models to commercialization projects, startups in the field of AI vision have recently been looking for business opportunities. At the AI Visual Creativity Competition hosted by the Shenzhen Longgang District People's Government, Shanghai Film, and Bilibili, exhibitor Wangou (Chengdu) Intelligent Technology Co., Ltd. COO Wen Xiaomeng told China Business News that his company is developing After deciding to enter the AI portrait project at the end of last year, we waited and watched for half a year. It was not until the second half of the year that we discovered that AI technology began to have the ability to be commercialized. Also applying AI to the commercialization process is Shanghai Fengjia Culture Communication Co., Ltd., an animation pre-development company. Yu Linfeng, co-founder of the company, told reporters that the company has been paying attention to AI for a year and a half. It was only in the past two or three months that it discovered that AI can be added to the animation process, so it figured out a set of workflows that greatly improved work efficiency. The commercialization of AI depends on the recent rapid maturity of AI technology. One creator described to reporters that AI technology has evolved from "fake at a glance" when depicting portraits a year or two ago to "hard to distinguish true from false" now. The AI Visual Creativity Competition also includes a creative forum and a workshop. During the competition, China Business News reporters communicated with a number of exhibitors, creators and industry experts and found that the industry is no longer most concerned about "what AI can do" , but how to maximize strengths and avoid weaknesses and use AI to make money. Startups began to use AI. At the beginning of this year, some AI videos began to appear on various video platforms. At that time, well-made AI videos were still in the minority, and creators were also waiting to see business opportunities. In the second half of the year, the rapid iteration of AI technology has led to a group of start-up companies willing to "try it out" and decide to enter the industry, and different business models have begun to emerge. This kind of technology iteration, in Wen Xiaomeng’s opinion, is the emergence of AI’s “commercial implementation capabilities”. He has been paying attention to AI technology for many years. From 2021 to the end of 2022, he concluded that at that time, AI tools to synthesize characters were still at the stage of "fake at first glance", such as disorganized facial features and uncoordinated limbs. Since then, the AI-synthesized portrait has gradually become more human-like, until it is 99% close to the photo taken by the camera. Since June this year, the company has invested money and run business. The initial investment has not yet been recouped, but the AI portrait project has delivered 7,000 orders to C-end consumer users, and the flow has quickly "turned up." The business model of Wenxiaomeng Company is to use AI models to develop portrait products of different sets and themes. Users place orders on the online platform and upload portrait data, or go to offline collection points to collect portrait data, and then use AI facial depth synthesis technology to A set of portrait photos can be generated. Wen Xiaomeng said that the business opportunity the company seizes is that studio photo products are relatively expensive, with a set costing thousands of yuan, while AI portrait products require almost no photographers, post-processing personnel and venue costs. After the initial research and development, the cost of using cloud computing AI to generate portraits is not high. It takes about 18 minutes to generate 50 customer photos, and the direct computing cost does not exceed 2 yuan. The cost is low, and a set of AI portraits can be priced below 300 yuan. Yu Linfeng has been waiting and watching for a long time. Recently, he finally decided that the technology is mature enough, so he recruited a group of technicians to dismantle the traditional animation process and add AI. "A large part of the industry teams are still waiting for mature products to come out, but we can't wait." Yu Linfeng told reporters. As an animation pre-development company, Yu Linfeng's company originally did not have a manpower-intensive mid-term team, but outsourced this part of the work. However, with the addition of AI technology, outsourcing has become a substitute. He told reporters that compared with leaving work to an outsourcing team, using AI can reduce a lot of costs. "There is definitely a gap of half." Efficiency improvements also reduce the labor costs corresponding to a single product. "Originally, one second of animation required drawing 12 pictures, but now we only need to draw the first and last frames, and the middle 10 pictures can be completed by AI." Yu Linfeng said that in the past, a one-minute two-dimensional video, six people's pictures, The team may have to work on it for a month, but now it only takes a small team of 10 people to make a 5-minute video for more than a month. Creators looking for business opportunities are also jumping on board. This AI visual creativity competition received 3,396 sets of images and video works. Qingshan, a member of Yuemo Studio, the creator of the award-winning video work "Dusty Melody", told reporters that the purpose of participating was to test the level of the work and gain access to industry resources. "The iteration of AI technology has a great impact on creation. Last year, the AI video was not clear enough and the picture lacked continuity, just like PPT. Now the video has a logical transition and the quality has improved a lot." Another member of Yuemo Studio Mo Shendu said. She admitted that she has not yet seen a very clear monetization method for AI videos, but considering the quality and cost advantages of AI videos, AI videos will have a monetization model. In contrast, previous video works required a lot of investment to achieve the quality of today's AI videos, but using AI does not require hiring actors or buying cameras. After replacing manpower, in addition to entrepreneurial teams that actively use AI tools to make money, AI tools have rapidly swept across all fields of visual creativity, making it difficult for practitioners in the traditional visual creativity field to avoid the influence of AI. The most obvious impact is the replacement of part of the workforce, as well as changes in industry pricing and practitioner work content. Yu Linfeng told reporters that the mid-term team may have more people, with a medium-sized team of about 80 people and a large team of more than 200 people. The traditional animation industry is reluctant to engage with AI in mid-term production. One factor is that the use of AI may reduce the team's manpower work. In addition, it is troublesome for large companies to adjust their processes, and they dare not make any changes until the technology is mature. Compared with human labor, AI will still bring obvious cost advantages. After AI technology matures, the cost of using AI and the cost of outsourcing related processes may be 1:10. Industry insiders told reporters that due to the impact of AI and market demand, outsourcing projects in the animation industry have indeed decreased this year. Luo Xiao, who is engaged in interactive media, artificial intelligence, and generative art creation and teaching, and is also the founder of experimental programming, told reporters that AI may not replace related work types that quickly, but it has replaced a lot of manual labor, such as poster production, Illustrations and game production processes use AI, and advertising companies also use AI when making proposals to clients. In the preliminary work of design, using AI becomes an established workflow. After AI penetrates into the workflow, creators must also adapt to different work requirements. "The most intuitive thing is the time requirement. In the past, projects that might have a month's working time have now been reduced to a week or even three days. This mainly affects design projects." Luo Xiao said that even if creators do not use AI, customers will follow the instructions. Based on the working time of AI to evaluate, this trend is gradually emerging. In addition, customers assume that creators save money due to AI, so the price of commercial projects is also driven down. "Creative ideas that originally took a week to realize may now need to be thought through in a day or half a day, and the rest is left to AI to quickly complete. While AI improves the overall efficiency, the creator's creativity and brain power increase in the entire creative process." "Concentration'." Luo Xiao told reporters that after using AI, creative work has become more demanding on the creator's mental work, and the focus of work is "compressed" on the mental work part, and the creator's mental consumption in a short period of time is much higher than before. As for whether creators can obtain higher returns than before after improving creative efficiency, it will be affected by business needs. "As for how AI can be monetized, we can't give a very precise answer now. With AI, everyone's overall level and requirements have become higher, but the product price will not necessarily become very high. And if everyone else is using the new Technology, if you are still doing it in the original way, you may be passively 'rolled'," Zhao Bozuo, executive chairman of the British International Film Festival Group (FFI) and master's tutor at Shantou University's Changjiang College of Art and Design, told reporters. Overall, the impact on creators will be that "the strong are stronger", and junior creators who rely too much on AI may end up with similar works. In the teaching process, AI is also replacing students’ work at hand. Zhao Bozuo told reporters that he is worried about this. As a university teacher, on the one hand, he hopes that students can use AI and keep pace with technological development. On the other hand, he is worried that students will use AI to simply generate works, turning AI into a search software. tools. How to maximize strengths and avoid weaknesses? New technologies are impacting traditional industries, and practitioners are in the process of adjusting their thinking and repositioning their careers. Zhao Bozuo told reporters that originally, art and design learners focused on practice and emphasized the ability at hand. Now it seems that beginners need to have a basic aesthetic appreciation and pay attention to the study of historical theories and genres that may have been ignored originally. These theoretical studies may become more important now. After AI made creation itself easier, many creators also expressed to reporters that aesthetics and creativity have become more important. "After work efficiency has improved, technical requirements have been reduced. Even without professional training, people can create their own works, but at the same time, the works may tend to be homogeneous. To create differentiated high-quality products, the core is creativity and aesthetics ." Qingshan thought. Wen Xiaomeng said that the leading AI visual generation companies in the industry do not have a cliff-edge advantage over each other. It is expected that in another year, the performance of technologies such as AI-generated images may hit the ceiling. By then, the model itself will no longer have major iterations, and the works produced by people using AI tools may be at the same level. “When everyone has a powerful tool and can generate a realistic portrait by just writing 10 words as prompts, the competition between companies will no longer be about whose model is better, but about aesthetics and creativity. . AI tools are only one of the elements of commercial companies, and the focus of competition in the future will return to the business itself," Wen Xiaomeng said. In the long run, the industry is still looking forward to AI bringing new vitality to the visual creative industry. Zhao Bozuo told reporters that he uses AI for product design, comic design, artwork, etc., which has opened up a lot of ideas. Taking the animation industry as an example, Yu Linfeng believes that as a mature market, the pain point of animation is that the production time is very long. In the mid-to-late stage, the production time of a 26-episode, 11-minute animation series basically takes 2 years, and the capital occupied is tens of millions of yuan, resulting in high risks. After using AI, the time can be shortened and the cost can be shortened. reduce. With the help of AI technology, a cartoon that originally cost 30 million yuan to produce may only cost 2 to 3 million yuan to complete, and the audience will be able to see more animations.
