With the recent developments in artificial intelligence, it has never been easier to fake or manipulate digital media like audio or images. As a consequence, more and more such content gets exposed on the Internet in social media, news, etc.
The Content Authenticity Initiative (CAI) develops a standard for verifying where media originates and whether it has been modified using Content Credentials. This helps users to check whether digital media can be trusted. In fact, this is already used in images generated with ChatGPT and is also supported in Adobe Photoshop. There are already libraries and tools available that can be used to implement it. While the standard mainly considers static data, there has also been a proposal on how to apply it to live videos.
The goal of this thesis is to investigate the use of Content Credentials and develop a real-world application where media content can be verified. Alternatively, it is also possible to look more closely into verifying data transmitted in real-time.
Does this sound interesting to you? Please get in touch with the supervisor!