Meta has announced guidelines for labeling, stating that images posted on Facebook, Instagram, and Threads are created with AI to prevent misunderstandings. This labeling will be implemented in the coming months. An example provided by Meta shows a label stating “AI Info” in a post (the actual text may vary in the future). This label will appear even on images created with AI from other companies such as Google, OpenAI, Microsoft, Adobe, Midjourney, and Shutterstock. These companies have agreed to include metadata and invisible watermarks in AI-generated images (for those interested, the standard metadata used is C2PA and IPTC).
Meta also mentions that the current metadata standard only applies to image media. There is no standard for sound and video media yet, making it difficult to check sound and video using AI tools from other companies. Currently, the approach is to have users manually provide information. As for industry agreements, they will be developed in collaboration with other players in the future.
TLDR: Meta announces guidelines for labeling AI-generated images on their platforms to prevent misunderstandings and ensure transparency. Other companies, such as Google and Microsoft, will also include metadata and invisible watermarks in their AI-generated images. However, there are no standard guidelines yet for sound and video media, requiring users to manually provide information. Industry collaborations will be established in the future to address this issue.
Leave a Comment