The rise of artificial intelligence (AI) has brought many benefits, but also many challenges. One of the most pressing issues is how to distinguish between real and fake content generated by AI, such as images, videos, and documents. To address this problem, a group of companies, including Microsoft, Adobe, and others, have created a new system to identify AI-generated content and its source. The system is called the Content Authenticity Initiative (CAI), and it consists of a symbol and metadata that can be added to any digital content.
The symbol is a “cr” bubble that stands for “content reality”. It is designed to be easily recognizable and unobtrusive. The symbol can be embedded in images, videos, and PDFs, and it can be clicked or tapped to reveal more information about the content. The information includes the origin, author, and modifications of the content, as well as the AI tools used to create or edit it. The metadata is stored in a secure and decentralized way using blockchain technology.
The CAI system is voluntary and aims to increase transparency and authenticity in the digital world. The creators of the system hope that it will help users to make informed decisions about the content they consume and share. They also hope that it will encourage ethical and responsible use of AI by content creators and editors.
However, the system is not without limitations and challenges. For one thing, it relies on the cooperation and honesty of the content creators and editors, who can choose whether to use the symbol or not, and what information to disclose or hide. For another thing, it does not prevent malicious actors from creating fake content without the symbol, or from tampering with the symbol or metadata. Moreover, it does not provide any verification or validation of the content or its source.
Therefore, the CAI system is not a foolproof solution to the problem of fake content. It is rather a tool that can help users to be more aware and critical of the content they encounter online. As one of the CAI partners said, “The ultimate goal is not to tell people what is real or fake, but to give them more information so they can decide for themselves”.
Relevant articles:
-Adobe created a symbol to encourage tagging AI-generated content, The Verge, Oct 10, 2023
-Microsoft, Adobe, and others launch new AI watermarking system, Wired UK, Oct 11, 2023
-How to spot fake images with the new AI watermarking symbol, Popular Science, Oct 13, 2023