Google, Bytedance face new IP lawsuits as AI copyright battles intensify
TL;DR
- 1Google est poursuivi par David Greene de NPR pour clonage de voix IA présumé dans NotebookLM.
- 2Hollywood accuse Seedance 2.0 de Bytedance de contrefaçon flagrante pour la reproduction de personnages et de voix.
- 3Des développeurs comme Google et OpenAI luttent contre les "attaques par distillation" qui clonent leurs modèles, tandis que les tribunaux remettent en question les droits d'auteur des œuvres purement générées par l'IA.
AI Copyright Clashes Escalate: Google, Bytedance Face Fresh Legal Challenges
The rapid evolution of artificial intelligence continues to test the boundaries of intellectual property and copyright law, leading to a surge in high-profile disputes. Tech giants Google and Bytedance are currently at the center of new legal battles, highlighting the multifaceted challenges emerging from AI’s ability to mimic and generate content.
One prominent case involves longtime NPR host David Greene suing Google. Greene alleges that the distinctive male podcast voice used in Google's NotebookLM tool is based on his own, raising critical questions about the unauthorized use of personal identity and voice. Simultaneously, Hollywood organizations are pushing back fiercely against Bytedance’s new Seedance 2.0 video generator. Dubbed a “virtual smash-and-grab” by critics, Seedance 2.0 is reportedly capable of generating highly realistic Disney characters, replicating actors' voices, and recreating entire fictional worlds, leading to accusations of “blatant” copyright infringement. These instances underscore the direct threat AI poses to creators whose works and identities can be replicated with unprecedented ease.
The IP challenges extend beyond direct content replication to the very models themselves. Google and OpenAI, despite facing their own lawsuits over training data, are now complaining about "distillation attacks" that systematically clone their billion-dollar AI models on the cheap. This phenomenon, alongside the competitive pricing of new models like Bytedance’s Seed2.0, which matches Western benchmarks at a fraction of the cost, intensifies the economic pressure on developers and raises concerns about the protection of foundational AI innovations. Meanwhile, the legal system struggles with defining authorship in the age of AI. A German district court recently denied copyright protection for three AI-generated logos, ruling that even elaborate prompting isn't sufficient for human authorship when the creative work is ultimately executed by the AI.
These developments paint a clear picture: the existing legal framework, built for a pre-AI era, is increasingly inadequate. From protecting individual voices and copyrighted characters to safeguarding the intellectual property embedded within complex AI models and determining authorship for AI-generated works, the challenges are profound. As AI tools become more sophisticated and accessible, the need for updated legislation and clearer ethical guidelines becomes paramount to ensure fair compensation for creators and foster responsible innovation in the AI landscape.
Sources
Weekly AI Newsletter
Trends, new tools, and exclusive analyses delivered weekly.