Hollywood Tames AI: ByteDance Signs Landmark Copyright Deal

TikTok owner ByteDance has officially agreed with Hollywood film studios to implement intellectual property protection rules within its artificial intelligence models. The Motion Picture Association (MPA) confirmed the signing of a memorandum that requires the AI developer to protect studio content from unauthorized use.

What the ByteDance and MPA Agreement Entails

These joint rules are aimed at curbing copyright infringements that could arise during image and video generation by ByteDance’s AI models. Industry giants such as Disney, Paramount, and Warner Bros. Discovery have joined the agreement. The MPA called this event a significant milestone in cooperation, noting that ByteDance’s latest models — Seedream 5.0 Pro and Seedance 2.5 — already demonstrate the implementation of necessary safety mechanisms.

Which Services Are Covered

The restrictions do not apply to individual tools but cover the company’s ecosystem of products on a global scale. The memorandum includes:

• All AI models used within the TikTok platform;
• The TikTok USDS joint venture service;
• The CapCut video editor;
• The Dreamina image generator.

Background: From Digital Twins to Memorandum

Previously, relations between the parties were extremely tense. In February, the MPA sent an official request to ByteDance demanding an end to the use of copyrighted materials after AI-generated deepfakes of Tom Cruise and Brad Pitt surfaced online. ByteDance responded promptly, promising to strengthen oversight and temporarily halting the global release of its Seedance 2.0 model. Charles Rivkin, CEO of the MPA, described subsequent negotiations as constructive, noting that the company has already integrated significant guardrails into its algorithms.

The deal between ByteDance and Hollywood sets an important precedent for the entire IT industry: major AI developers are being forced to move away from the unchecked use of third-party content and toward legal operational frameworks. This decision will not only reduce the risks of deepfakes on social media but will also set a new standard for the regulation of generative neural networks worldwide.