Seedance 2.0 signals big shift in AI sector

Chinese tech company Byte-Dance has launched the latest version of its artificial intelligence video-generation model Seedance 2.0, intensifying China's push in generative AI at a time when the United States' OpenAI appears to have slowed development of its similar product, Sora.

Seedance 2.0, which is currently in limited testing, is able to generate a multi-shot film sequence with sound in roughly 60 seconds. The model has attracted attention from developers and on social media, where users have been sharing clips showing cinematic scenes, fluid camera movements and synchronized audio — all generated from simple prompts.

Feng Ji, CEO of Game Science, the studio behind the video game sensation Black Myth: Wukong, described the system as a game-killer. He said the technology marked the end of the "childhood" phase of AI-generated content, or AIGC.

Industry experts said ByteDance's latest move underscores how Chinese companies are racing to gain an edge in the AIGC sector — a field with huge commercial potential across the entertainment, advertising, gaming and social media segments.

While OpenAI's text-to-video model Sora set off global excitement upon its preview, public updates on its broader rollout have been limited in recent months, leaving room for competitors to gain visibility, experts said.

Feng said that Seedance 2.0 and the technology behind it could trigger a structural shift in the video industry and an explosion of AIGC offerings.

"The cost of producing ordinary videos will no longer follow the traditional logic of the film and television industry, and will instead begin to become increasingly low. This would compel studios and platforms to rethink workflows that have long depended on large crews and expensive equipment," he added.

Pan Helin, a member of an expert committee under China's Ministry of Industry and Information Technology, said that Seedance's unexpected performance can be partially attributed to ByteDance's vast content ecosystem, as the company operates some of the world's largest short-video platforms with access to extensive data on visual styles and user preferences.

"Compared with some overseas competitors, the new model can better help people create short videos. Such alignment could make the model particularly attractive to online creators in both China and abroad," he said.

Investors have been quick to respond. Shares of several Chinese media and technology companies associated with AI content production and Byte-Dance-related companies have surged in recent trading sessions.

However, some concerns remain about the rapid progress of AIGC technology and related products. Some early users said that after uploading just one photo, the system is able to generate voices closely resembling real individuals — even without audio samples — posing a data security concern.

Sha Lei, a professor at Beihang University's AI research institute, said: "Most large AI models, both in China and abroad, are trained on publicly available data. Authorization and boundaries of use are still being explored."

"Also, increasingly realistic AI-generated videos could intensify data security problems like deepfakes. Regulators, platforms and creators are likely to face growing pressure to balance innovation with safeguards," Sha said.