The tech race between big companies to develop the most advanced AI models continues with the release of ByteDance’s new video generator. In a post, ByteDance – the company behind TikTok – says that the new Seedance 2.0 supports commands that combine text, images, video, and audio.
The company claims it offers a significant improvement in the quality of creation, offering improvements in the creation of complex scenes with multiple subjects and in its ability to follow instructions. Users can improve their text commands by providing Seedance 2.0 with up to nine images, three video clips and three audio clips.
See also: TikTok in the US introduces 'Local Feed'
The model can create clips up to 15 seconds long with audio, taking into account camera movement, visual effects and motion. It can also reference text-based storyboards, according to ByteDance.
Explaining how its model works, ByteDance shared an example, showing two figure skaters. The company says Seedance 2.0 can “reliably execute a sequence of highly challenging moves — including synchronized takeoffs, midair spins, and precise landings on the ice — while strictly following the physical laws of the real world.”
Users on social media have already started showing off what the new tool can do, with one person posting an AI video of two characters resembling Brad Pitt and Tom Cruise in a movie fight scene.
See also: EU threatens TikTok with fine over 'addictive design'
Other posts show Seedance 2.0's ability to create anime-style clips, cartoons, sci-fi movie scenes, and videos that look like they were created by content creators. It's unclear what (if any) copyright protections Seedance 2.0 offers, as a quick search on X reveals a number of clips featuring characters from Dragon Ball Z, Family Guy, Pokémon, and more.
For now, Seedance 2.0 is only available through Dreamina AI and its AI assistant, Doubao. It's unclear whether it will make it to TikTok — especially now that the app is under new ownership in the US.
See also: USA: Who's afraid of the new TikTok?
ByteDance Seedance 2.0 and competing AI models
AI-powered generation models have become more advanced in the past year, with Google Veo 3 adding the ability to create clips with audio support and OpenAI releasing Sora 2 along with a new app that lets users create videos with “hyper-realistic motion and sound.” AI startup Runwayhas also released a new version of its AI model, which it claims has “unprecedented” accuracy.
