字节又一AI产品刷屏 网红博主担忧“被训练”

Core Viewpoint - The launch of ByteDance's Seedance 2.0 video generation model has sparked significant discussion in the AI industry, highlighting both its advanced capabilities and the ongoing concerns regarding copyright issues related to AI training data [1][4]. Group 1: Seedance 2.0 Model Capabilities - Seedance 2.0 supports various input types including text, images, videos, and audio, enabling the generation of videos with advanced features such as self-rotation, multi-modal thinking, audio-visual synchronization, and multi-angle storytelling [1]. - AI creator feedback on Seedance 2.0 has been largely positive, noting a significant improvement in video quality compared to previous AI models [2]. Group 2: Copyright Concerns - The model's ability to generate content using personal attributes, such as matching a user's voice without explicit permission, raises serious copyright and ethical questions [2][3]. - The ongoing debate about copyright in AI training data is not limited to ByteDance; other companies like Anthropic, OpenAI, and Stability AI are also facing similar issues regarding the use of publicly available data for training [4]. - Legal experts indicate that the current legal framework regarding AI training data and copyright is still ambiguous, with no definitive rulings in either China or the U.S. [5][6]. Group 3: Legal Implications - The lack of clear legal guidelines on AI-generated content and copyright issues is causing courts to pause on related cases, as AI is considered a new phenomenon [5]. - There are suggestions for collective copyright management mechanisms to address industry-wide copyright disputes, although practical implementation remains uncertain [5]. - Specific legal protections exist for voice and image rights, indicating that AI-generated content that uses personal attributes without consent could infringe on these rights [6].