Chinese tech company ByteDance announced Seedance 2.5, the latest version of its artificial intelligence video generation model, at a conference in Beijing on Tuesday, The Information reported.
AI video generation has come a long way since its debut, with each new version getting better and better at producing realistic images. It’s a far cry from the first time I saw Will Smith eating spaghetti, but it was horribly bad.
Now, these AI-generated videos need to be watermarked to make it easier to identify deepfakes and other things. Synthetic or false content.
The latest version of the video model allows users to provide up to 50 reference files (up from 12 in previous versions), whether images, videos, or audio files. Seadance 2.0. Increasing the number of references gives you more control over the video creation process. This model can generate 30 seconds of 4K video with a single prompt.
ByteDance has consistently released some of the most impressive AI video generation models, including: OpenAI’s now defunct Sora and Google Veo3. According to reports, ByteDance, which previously held a majority stake in TikTok, is said to be launching a new model in China next month. There is no mention of release dates in other countries.
Introducing a new model may attract some people’s attention, but it’s not the best way to go. Seedance 2.0 rollout in the US delayed Earlier this year, it came under pressure from Hollywood to stop using copyrighted works that were supposedly used to train models. Even if the latest model is significantly better than its predecessor, a similar backlash could occur if legal and copyright issues are not addressed.
ByteDance did not respond to a request for comment.
