The MPA urges ByteDance to curb its AI video model Seedance 2.0, after an AI-generated video of Tom Cruise and Brad Pitt fighting on a rooftop went viral
The dispute matters because Seedance’s realism, low pricing and timeline-based prompting were positioned as routes into Hollywood-facing creation. Subsequent heavy usage also exposed a practical constraint: demand strained Seedance 2.0’s compute capacity, even as legal pressure centered on what the model can generate and how it is trained.
First-order effects
ByteDance faces immediate pressure to add or tighten safeguards around copyrighted characters, celebrity likenesses and other film-related outputs from Seedance 2.0.
The MPA’s intervention gives member studios a coordinated industry vehicle for challenging the tool, rather than leaving enforcement solely to individual rights holders.
Second-order effects
Studios and other media owners gain a stronger basis to seek comparable restrictions from competing AI-video providers, making rights-management features a more important product and partnership issue.
For ByteDance, the combination of demand-related capacity limits and content-control demands could complicate rollout priorities: scaling access without expanding allegedly infringing use becomes harder.
Third-order effects
If coordinated studio pressure persists, AI-video competition will increasingly turn on controllable provenance, training-data governance and output restrictions—not only visual quality or price.
The episode points toward a synthetic-media control layer in which entertainment IP owners seek recurring influence over model training and generation policies; the eventual balance will depend on legal and commercial settlements that are not established here.
The trend: Generative-video platforms are moving from viral consumer products toward tightly governed media infrastructure shaped by IP owners, capacity constraints and access controls.
Seedance 2.0 just dropped from ByteDance Seed team and this is a MASSIVE upgrade for video generation 🧵 Unified multimodal audio-video joint generation. You can feed it: • Up to 9 images • 3 video clips • 3 audio clips • + text instructions All at once. Here's what it can [video]
Evals show Seedance 2.0 leading across T2V, I2V, and multimodal reference tasks. Available NOW on Dreamina AI and Doubao. Another Chinese lab shipping absolute bangers. 🔗 Homepage: https://seed.bytedance.com/... 📝 Blog: https://seed.bytedance.com/... Follow @altryne and @ThursdAI…
Seedance 2.0 Prompt: just toss a bunch of bullshit on screen, show me like a big ship too, everything fucking blows up - make sure its insane and gets at least 50 likes [video]
This video of Brad Pitt fighting Tom Cruise and angry about killing Epstein looks completely real. It's not. It's AI. ByteDance's Seedance 2.0 generates this in minutes. Think about what this means for the next election. [video]
⚡️There is no going back. What's happening here is a terminal shift in how humans interface with reality. When anyone can generate footage of anything and the outputs feel real in the body, the boundary between witnessing and imagining disappears. This is a perceptual severing.
Not to be a pedant but this particular clip being the focus is weird considering it's *not* infringing any particular work. It's just using actors faces.