Google sued over AI voice; Hollywood warns on Bytedance's Seedance 2.0
TL;DR
- 1L'animateur de NPR, David Greene, poursuit Google pour clonage de voix non autorisé présumé dans son outil IA NotebookLM.
- 2Hollywood lutte contre Seedance 2.0 de Bytedance, le qualifiant de « vol à l'arraché virtuel » pour sa capacité à reproduire des personnages et des voix protégés par le droit d'auteur.
- 3Un tribunal allemand a refusé le droit d'auteur pour des logos générés par IA, exigeant la créativité humaine, ce qui ajoute aux incertitudes juridiques concernant les productions d'IA.
The generative AI landscape is increasingly entangled in complex legal battles, as creators and industries confront alleged copyright infringement and unauthorized use of their work. Recent developments highlight the urgent need for updated legal frameworks to address the rapid advancements in AI technology.
Longtime NPR host David Greene has launched a lawsuit against Google, asserting that the distinctive male podcast voice utilized in the company's NotebookLM tool is based on his own voice without consent or compensation. This case underscores a growing concern among creators about AI models replicating their likeness or voice. Simultaneously, Hollywood organizations are sounding the alarm over Bytedance's new Seedance 2.0 video generator. They allege the model has become a primary vehicle for "blatant" copyright infringement, capable of replicating famous Disney characters, cloning actors' voices, and recreating entire fictional worlds with startling realism, which some describe as a "virtual smash-and-grab" against intellectual property. (Source: TechCrunch AI) (Source: TechCrunch AI) (Source: The Decoder)
The capabilities of models like Seedance 2.0 are not only causing legal headaches but also exerting significant market pressure. Bytedance's new series of AI models are reportedly matching Western competitors in benchmarks while being offered at a fraction of the cost. The accessibility of sophisticated generative audio tools is also growing, with open-source models like 'Kani-TTS-2' demonstrating high-fidelity speech synthesis and voice cloning capabilities that run on modest hardware, further democratizing access to potentially contentious technologies. (Source: The Decoder) (Source: MarkTechPost)
Adding another layer to the complex legal landscape, a German district court recently denied copyright protection for three AI-generated logos. The ruling stated that even elaborate prompting is insufficient to grant copyright when the ultimate creative work is left to the AI, highlighting a foundational debate on what constitutes 'human creativity' in the age of generative AI. These ongoing cases and rulings underscore a critical juncture where existing copyright law, built for a pre-AI era, struggles to keep pace with technological advancements, necessitating a global conversation on legal reform and ethical guidelines. (Source: The Decoder)
Sources
Weekly AI Newsletter
Trends, new tools, and exclusive analyses delivered weekly.