ByteDance, the parent company of TikTok, has unveiled the pre-release version of its new AI video model, Seedance 2.0, which has already begun to impact the stock market by boosting the shares of Chinese AI companies. Launched over the weekend, Seedance 2.0 positions itself as a “true” multi-modal AI creator, capable of merging images, videos, audio, and text to produce what it describes as “cinematic content.” With advanced features like precise reference capabilities, seamless video extension, and natural language control, the model is now accessible to select users of Jimeng AI, ByteDance’s AI video platform.
One of the standout improvements in Seedance 2.0 compared to its predecessor, version 1.5, is its ability to export videos in 2k resolution, along with a generation speed that is 30% faster. An evaluation from Swiss-based consultancy CTOL has labeled Seedance 2.0 as the “most advanced AI video generation model available,” asserting that it has outperformed competitors such as OpenAI’s Sora 2 and Google’s Veo 3.1 in practical assessments.
The enthusiastic reception of Seedance 2.0 has positively influenced the market, with shares of the publishing company COL Group Co reaching its daily price ceiling of 20%. Likewise, Shanghai Film Co and gaming and entertainment firm Perfect World Co saw their shares increase by 10%. Notably, the Shanghai Shenzhen CSI 300 Index has risen by 1.63% as of the latest updates.
The rapid advancement of consumer AI video generators has made noticeable quality improvements within a short timeframe. While previous AI-generated videos often struggled with common issues like blurry fingers and unrealistic skin textures, these flaws are becoming increasingly difficult to detect. In contrast to competing models such as Sora 2, which includes watermarks in its output, and Veo 3.1, which features a metadata watermark called SynthID, Seedance advertises its products as completely watermark-free.
However, the rise of sophisticated AI tools and their accessibility has raised concerns regarding the potential for creating deepfake content. This worry has escalated, especially following last month’s EU investigation into X (formerly Twitter), examining whether the social media platform effectively managed the risks associated with its AI chatbot Grok, which was equipped to edit images. Reports indicated troubling uses of Grok, including the capability to undress figures in images, leading to the generation of millions of inappropriate content pieces.
As AI technology continues to evolve, the implications for both creativity and ethical usage are significant. The promising features of Seedance 2.0 present exciting possibilities for content creation while highlighting the importance of responsible AI development and oversight in the digital landscape.