ET Explained: AI race in focus as ByteDance hits play on AI video creator
ByteDance last week launched a new text-to-video generative AI app called Jimeng AI, said to be a rival to OpenAI’s Sora which is yet to be made publicly available. ET takes a close look at Jimeng and other Chinese GenAI models, and their place in the global AI race.
What does Jimeng AI do?
Jimeng AI was developed by ByteDance-owned Faceu Technology and can generate images or videos from text prompts. It is available on both Android and Apple’s App Stores, but only in China.
The Chinese word Jimeng translates into “immediate dream”.
Users can use the app to create 80 images or 26 videos for free. The app has monthly subscriptions starting at 69 yuan or about $10 to generate more videos.
Mainland China’s AI video-generation market is projected to reach 9.3 billion yuan by 2026, up from 8 million yuan in 2021, according to research firm LeadLeo.
OpenAI’s Sora remains available only to red teamers to assess critical areas for harms or risks, and visual artists, designers and filmmakers for feedback.
Has China produced other prominent GenAI models?
China has seen significant domestic competition in the video models arena. Last month, another short-video platform, Kuaishou, launched text-to-video model Kling AI, whose beta version is accessible worldwide for those who register with email.
On the heels of this launch, AI startup Zhipu introduced its text and image-to-video model Ying. The startup also has a chatbot called ChatGLM.
Similarly, Tsinghua University and AI startup ShengShu Technology created Vidu AI,
