全球知名的人工智能研究实验室OpenAI近日宣布，其先进的图像生成工具DALL-E 3将开始集成水印功能，旨在帮助用户辨别由人工智能生成的图像。这一举措是OpenAI对内容真实性和来源透明度的进一步承诺，也是对当前数字媒体时代中AI创作日益普及的回应。
OpenAI表示，新添加的水印将基于内容来源和真实性联盟（C2PA）的标准，旨在确保图像的可追溯性和可信度。水印将出现在ChatGPT网站以及通过DALL-E 3模型API创建的图像上。值得注意的是，移动设备用户预计从2月12日起将开始看到这些水印。
水印系统由两部分组成：一个不可见的元数据组件，用于隐藏式记录图像的生成信息；另一个是可见的CR（创建的现实，Created Reality）符号，该符号将被置于每个生成图像的左上角，直观地提示观众该图像的生成方式。
这一更新对于新闻机构、艺术家、版权持有者以及广大互联网用户来说具有重要意义，因为它有助于防止误导性的AI生成内容被误用或冒充为原创作品。OpenAI的这一创新举措不仅加强了其在AI伦理和责任方面的领导地位，也预示着未来数字内容验证技术的发展方向。
英语如下：
**News Title:** “OpenAI’s DALL-E 3 Introduces Watermark Feature to Authenticate AI-Generated Images”
**Keywords:** OpenAI, DALL-E 3, AI Watermark
**News Content:**
### OpenAI’s DALL-E 3 Image Generator Adds Watermarks to Aid AI-Created Image Identification
OpenAI, the renowned artificial intelligence research laboratory, recently announced that its advanced image generation tool, DALL-E 3, will incorporate watermark functionality to assist users in distinguishing images created by AI. This move represents OpenAI’s further commitment to content authenticity and transparency in source and responds to the growing prevalence of AI-generated content in today’s digital media landscape.
According to OpenAI, the new watermarks will adhere to the standards set by the Content Authenticity Initiative (C2PA), ensuring the traceability and credibility of images. These watermarks will appear on images generated through the ChatGPT website as well as those created via the DALL-E 3 model API. Notably, mobile users are expected to start seeing these watermarks from February 12th.
The watermark system consists of two parts: an invisible metadata component that discreetly records the image’s generation information; and a visible “CR” (Created Reality) symbol placed in the top-left corner of each generated image, indicating to viewers the artificial nature of the image.
This update holds significant implications for news organizations, artists, copyright holders, and general internet users, as it helps prevent misleading AI-generated content from being misused or masqueraded as original works. OpenAI’s innovative step reinforces its leadership in AI ethics and responsibility and foreshadows the direction of future digital content verification technologies.
【来源】https://www.ithome.com/0/749/371.htm
Views: 1