The Kling AI platform of Kuaishou generates a video from text and fixed images.
Nurphoto | Nurphoto | Getty images
Beijing – The world of Chinese video entertainment has given a data mine for companies – and they are now accelerating winning artificial intelligence tools to generate cinema advertisements and clips.
Tiktok Parent Bytedance holds the first and third places of the research company Artificial analysis“The leading generative texts in the leading video text, which have been launched in the past two months. Google holds the second and fourth places, while the short video application based on Beijing Kuaishou’s Kling Ai ranks fifth.
Despite a certain consolidation in other parts of the AI industry, “competition in AI videos generation models is at a previous stage, and some Chinese companies have emerged as the first leaders in this space,” said Wei Xiong, Internet analyst in China at UBS Securities.
“We believe that the generation of AI videos has the potential to reshape the content industry,” she said, “improving production efficiency, reducing obstacles to creation and unlocking new models of monetization.”
With such AI tools, users can download a single image or more and direct AI to generate a video clip based on them. Other tools allow users to enter text, from which AI will generate the video clip.
More than 20,000 companies, advertisers to film leaders already use Kling AI to generate a video, the company based in Beijing said this week at the IA World Conference in Shanghai. The latest version, Kling 2.1, can automatically add relevant sound effects to match the video generated by AI.
It is not only for users in China.
“Whether it is the user scale or commercial income, abroad, explains the majority,” said Zeng Yushen, operations manager at Kling AI. She said the company planned to improve its support for the tool in places like Japan, South Korea and Europe.
“This is something we have observed, the great models of AI are increasingly globalized,” she said. “People do not seem to worry about the country’s product.”
Kuaishou claimed Kling Ai more than 150 million yuan ($ 20.83 million) in income During the first three months of the year, and that daily advertising spends in generative AI tools was 30 million yuan during this period. The company has not yet announced when it will publish the results of the second quarter. Zeng refused to share the training costs of Kling AI models.
Although the cost of reduced production implies a “size” market, Xiong of UBS said: “The current capacity of the model remain limited by the length of the clips, the coherence of the movement and the controllability”.
AI Chinese companies are also faced with competition from the United States, beyond the Trump administration restrictions on China access to advanced semiconductors necessary for the training of AI models.
Amazon And google has Tools launched to generate a video from images or text. Outings come as Microsoft-Abosque Openai launched its GENERATION MODEL OF VIDEOS SORA to Chatgpt subscribers in December – almost a year after revealing its capacities in February 2024.
However, Kling AI had already launched the public in June 2024. Users subscribe and bought credits to generate videos.
Vidu, a startup rival tool based in Beijing, Shengshu, was launched among world users about 12 months ago, and around this year, it expected it to expect an annual turnover of $ 20 million on the basis of user subscription costs.
“Chinese companies tend to try first to identify a commercial” pain “…, areas where companies will pay for services, which has been a challenge for AI requests,” said Paul Triolo, partner and vice-president of China in the DGA-Albright Stonebridge Group consulting firm.
He underlined how the Chinese Startup 3DSTYLE uses a generative AI to design new styles of clothing and integrate them into automated manufacturing connected to the Internet.
US companies have also applied AI to specific industries, Triolo said, but Chinese companies are often able to integrate AI faster because they are faced with a very competitive environment and can recruit in a “highly qualified” local engineers.
‘Ai as a filmmaker’
Chinese electronic commerce giant Alibaba also remained at the top of the trend by publishing the latest version of its video generation model this week called WAN2.2. The company said that with the open source model, users can control the lighting, the time of day, the color tone, the angle of the camera, the size of the frame, the composition and the focal distance.
Open Source allows users to download a model for free and personalize, if not market, products with it. Alibaba said that since the open supply of the series of “Wan” models in February, the models have been downloaded more than 5.4 million times from the grip and similar facial platform in China called Modelcope.
“The age of AI in the cinema is over. We have entered the age of AI as a filmmaker,” said Winston Ma, auxiliary professor at the Nyu School of Law. He pointed out that 1.4 billion people from China gave local businesses “huge” data of data with video data with which to work.
“Just as Tiktok has taken over the global markets with short videos in the Mobile Internet era, Chinese AI companies could well direct the generative AI revolution in visual digital entertainment,” said MA, author of “The Digital War: How China’s Tech Power shapes the future of AI, blockchain and cyberspace”.
Avatars and games
Chinese companies also build AI tools for more than the simple generation of videos.
During last week, Baidu announced that its new digital human technology fueled by the AI – which propelled sales of $ 7.65 million during an interactive live broadcast session of more than six hours in June – would be published for wider use in October.
In 3D visualization, Tence Published its HUNYUAN World model to create digital panoramic images of scenes, generated from text and visual prompts. Visuals use a “mesh” file format that players can then use to modify specific image parts.
“Beyond support [Tencent’s] The internal development teams, the platform demonstrate Tencent’s ambition to standardize the generation of high-fidelity game assets and to expand its influence through the landscape of Chinese game development, “said Daniel Ahmad, director of research and information at Niko Partners.
Niko found that more than half of game development studios in China already use AI for content generation and the reduction in time and development costs.
But the development of the game reflects wider challenges in the use of large -scale AI to generate videos and graphics.
“Although the interest in AI is high,” said Ahmad, “we have already seen reactions to games that have poorly implemented technology.”
