Pictured here is an AI-generated clip from Vidu’s website. The tool can create videos from text or images.
Evelyn Cheng CNBC
BEIJING – Beijing-based Shengshu Technology on Wednesday said its artificial intelligence-powered text-to-video tool Vidu will now be able to generate videos by combining multiple images.
Vidu has enabled users around the world to create 8-second clips based on written prompts. While OpenAI – the provider of ChatGPT – announced in February that Sora’s AI model can generate a one-minute video from text, it has not yet been publicly released.
Vidu’s new AI feature can combine three images – such as a shirt, a person and a moped – into a video of a person wearing a shirt and driving a moped through a scene, Shengshu said.
Other platforms claim to be able to convert text or images into videos using AI, but the output quality varies. Shengshu’s claimed breakthrough is the ability to take three unique images and combine them with visual consistency into an AI-generated video.
“Very early on, we identified (visual consistency) as a problem, and wanted to solve it well,” said Fan Bao, Shengshu’s chief technology officer, in Mandarin, translated by CNBC.
Vidu launched in April and its ability to turn two profile photos into a video like people hugging went viral on TikTok.
The AI ​​video generator has earned money from advertisers, animators and other businesses, co-founder and CEO Shengshu Jiayu Tang said in Mandarin, according to a CNBC translation. He said the monthly usage rate per customer can range from 100,000 yuan to 1 million yuan ($13,871 to $138,711).
To overcome copyright issues, Tang said the company could sign deals with artists that would allow AI to imitate the artist’s painting style for advertising. He said he had not seen any significant legal cases regarding the use of consumer images.
Tang added that Vidu does not allow the public to generate content using images of celebrities or “sensitive” individuals. He said the AI ​​tool also banned mute and violent images. As for personal photos, Tang said Vidu destroyed the data according to general data protection regulations – a global benchmark.
Shengshu was founded last year with backers including Baidu Ventures, Alibaba affiliate Ant Group, Chinese AI startup Zhipu, Qiming Venture Partners and the city of Beijing, according to PitchBook.
Tang said AI Vidu runs rented cloud servers in China and overseas.
