🧠 Food for thought

Implications, context, and why it matters.

Seedance’s range brings power and risk

• Seedance 2.0 goes beyond text-only rivals. It is described as a “quad-modal” system that takes text, images, videos, and audio at the same time for tighter control 1.
• It runs on a dual-branch diffusion transformer architecture (a type of AI model design for generating media). The setup creates video visuals and native audio together, which supports lip-sync and timed sound effects 21.
• ByteDance previously paused a feature after reports said it could recreate highly accurate personal voice characteristics from a facial photo. That sparked privacy and ethics questions 2.
• Pressure from entertainment groups has grown. The Motion Picture Association (MPA) urged the platform to stop what it described as infringement tied to unauthorized copyrighted content, while SAG-AFTRA (the actors’ and performers’ union) condemned what it described as unauthorized use of members’ voices and likenesses 34.

Hollywood is pushing a choice between licensed and unlicensed AI

• Studios and unions are treating licensed training and output as the line that separates acceptable AI from the rest, which puts new strain on the generative AI market.
• Disney sent ByteDance a cease-and-desist letter that accuses Seedance 2.0 of “hijacking” trademarked characters and pulling from a “pirated library” of Disney assets. Disney has also been reported to have put $1 billion into OpenAI and to be planning character licensing for OpenAI’s Sora video platform 3.
• After infringement worries around Sora 2, the Motion Picture Association said OpenAI added safeguards that make it harder to break studio copyright rules. The same source says Disney later agreed to license 200 characters for use on Sora 2 5.
• ByteDance did not immediately respond to requests for comment tied to the MPA criticism 6. Its next move could affect whether Seedance seeks deals with rightsholders or runs into more resistance in Western markets 6.