Shengshu AI says its Sora-like text-to-video tool will bring down production costs

Shengshu AI, an artificial intelligence (AI) firm focusing on text-to-video models, held an event on Wednesday dedicated to showcasing how its video-generation tool that rivals OpenAI’s Sora can maintain consistency and bring down costs in producing films and anime.

The event in Beijing, where Shengshu is based, focused solely on the “reference to video” feature in its Vidu tool, which company co-founder and CEO Tang Jiayu said maintains consistency of “subjects, scenarios and style” across clips, something that has been a challenge for generative AI. This will cut down on production time and save on costs, according to Tang.

The feature has been available in Vidu since it was publicly released in July. But the latest event comes amid rising competition in AI video production as companies race to find more commercial uses of models that are costly to train.

Short-video giant Kuaishou Technology and state-backed Zhipu AI have both launched similar video-generating tools – Kling and Ying, respectively.

Kuaishou advocated the adoption of Kling in the film industry earlier this week by having nine famous Chinese directors use it to make short films.