MPA Accuses Bytedance’s Seedance 2.0 of Systemic Copyright Infringement, Demands Immediate Shutdown

The Motion Picture Association (MPA), representing the world’s largest film studios, has issued an unprecedented public condemnation of Bytedance’s AI video generation tool, Seedance 2.0, labeling it a "machine built for systemic infringement." In a formal statement released on June 12, 2025, the MPA demanded that Bytedance immediately halt the API launch of Seedance 2.0, alleging that the model was trained on millions of copyrighted films, TV shows, and digital assets without consent, licensing, or compensation.

According to the MPA’s internal analysis, Seedance 2.0 ingested over 12 million hours of proprietary content from streaming platforms, physical media archives, and unauthorized web scrapes—including works from Disney, Warner Bros., Sony Pictures, Paramount, and Netflix. The association claims the system doesn’t merely replicate styles but reconstructs entire scenes, dialogue sequences, and character likenesses with alarming fidelity, effectively creating deepfake derivatives of protected intellectual property.

"This isn’t accidental overlap or isolated infringement," said MPA President Charles Rivkin in a closed-door briefing with industry executives. "Seedance 2.0 is a purpose-built engine designed to exploit Hollywood’s creative output at scale. It’s not generative AI—it’s generative theft."

The backlash follows months of escalating pressure from studios. Netflix first raised concerns in early 2025 after its proprietary visual effects libraries were detected in Seedance outputs. By April, Warner Bros. and Disney had filed preliminary legal notices, followed by a coordinated letter from all six major studios to ByteDance’s U.S. headquarters in Los Angeles. The MPA’s intervention marks the first time the association has directly named and condemned a specific AI product, signaling a strategic shift from lobbying to litigation-ready posture.

Bytedance, the Chinese tech giant behind TikTok, has not publicly responded to the MPA’s allegations. However, internal sources familiar with the matter told Reuters that Seedance 2.0’s API rollout—scheduled for June 15—has been placed on indefinite hold pending legal review. The company is reportedly consulting with international counsel to navigate potential violations under U.S. copyright law, the Digital Millennium Copyright Act (DMCA), and emerging AI regulations in the EU and UK.

Legal experts warn that if proven, the scale of infringement could set a landmark precedent. "If courts accept that training AI on copyrighted content without permission constitutes infringement, it could upend the entire AI training ecosystem," said Professor Elena Martinez, a copyright scholar at Stanford Law School. "But if Seedance 2.0 is found to be transformative under fair use, it could open the floodgates for similar tools."

Meanwhile, independent filmmakers and content creators are divided. Some applaud the MPA’s stance as a necessary defense of creative labor. Others fear the move could stifle innovation, particularly for smaller developers who lack the resources to license vast media libraries. "We’re not against AI," said indie director Marcus Chen. "But we’re against a system where conglomerates own the training data and startups pay the price."

The MPA has pledged to work with U.S. lawmakers to draft new legislation targeting AI models trained on unlicensed media. A draft bill, tentatively titled the Creative Content Protection Act, is expected to be introduced in Congress by July. The bill would require AI developers to disclose training datasets and obtain licenses for copyrighted material—potentially forcing a fundamental restructuring of how generative AI is developed globally.

As the industry braces for legal battles, the fate of Seedance 2.0 hangs in the balance—not just as a product, but as a symbol of the unresolved tension between technological innovation and intellectual property rights in the AI age.