OpenAI公司日前宣布，其最新的图像生成器DALL-E 3将引入一项新功能——添加来自内容来源和真实性联盟（C2PA）的水印，以帮助用户识别使用人工智能生成的内容。这一举措标志着人工智能技术在内容创作领域的一大进步，同时也为打击虚假信息和保护知识产权提供了新的解决方案。
DALL-E 3是OpenAI开发的一种高级图像生成模型，它能够根据文本描述创造出惊人的图像。随着人工智能技术的进步，AI生成内容的分辨和鉴别变得越来越重要。为了解决这一问题，OpenAI决定在DALL-E 3生成的图像中加入水印，以提高内容的透明度和可信度。
水印将包含两个部分：一个不可见的元数据组件和一个位于图像左上角的可见CR符号。这一设计旨在确保无论图像如何被编辑或分享，其来源和AI生成的事实都能够被识别。
移动端用户将从2月12日开始看到这些水印。这一新功能的推出，不仅是DALL-E 3的重大升级，也是AI内容领域的一大创新。OpenAI的这一举措，无疑将为人工智能在媒体和艺术领域的应用树立新的标准。
英文标题：OpenAI Introduces Watermark Feature to Identify AI-Generated Content in DALL-E 3
英文关键词：OpenAI, DALL-E 3, Watermark, AI-Generated Imagery, C2PA
英文新闻内容：
OpenAI, a leading AI research and deployment company, has recently announced a new feature for its advanced image generator, DALL-E 3. This feature involves the integration of content-source and C2PA (Content Authenticity Initiative) watermarks to help users identify images created using artificial intelligence. The move is a significant step forward in the transparency and accountability of AI-generated content, addressing growing concerns over misinformation and copyright protection.
The watermark will consist of two components: an invisible metadata layer and a visible “CR” symbol in the upper-left corner of each image. This design ensures that the origin and AI-generated nature of the content remain identifiable, even if the image is edited or shared.
Mobile users will begin to see these watermarks from February 12. The introduction of this feature marks a major upgrade for DALL-E 3 and a step towards standardization in the application of AI in media and artistic creation. OpenAI’s initiative is likely to set a new benchmark for the use of AI in content creation.
【来源】https://www.ithome.com/0/749/371.htm
Views: 1