As videos closely mimic reality, it is important to provide a speedy response on (a) whether a video is manipulated, (b) correct context and content on the video. It is also important to educate the public on the possibilities and limits of the technology in manipulating videos, and ways in which they can protect themselves. Finally, since the capabilities of AI in manipulating and generating videos is rapidly evolving, there is a need for data that enables research on this.
Rapid advances in generative AI has led to the emergence of AI generated content that is blurring the lines between fiction and reality. Fact checkers now don’t just have to point out what is false but also inform people what is indeed true. Over the past few months, a series of audio clips and videos that were altered (and in some cases, also real) has gotten the moniker “deepfakes”. To tackle the problem effectively there is a need for research on the types and challenges of deepfakes, as well as timely response.
The Misinformation Combat Alliance, has set up a Deepfakes Analysis Unit (DAU) to respond to and research content that comes on a WhatsApp tipline. DAU will coordinate with member fact-checking organisations, industry partners and digital labs.
This is the project management repository of the DAU Tipline.
🗄️ Three other repositories form the core components of the tipline:
| Repository | Description |
|---|---|
| Dashboard | A collaborative space for analysing deepfakes, this repository contains the frontend and backend code for facilitating interactions and workflows for this analyses. |
| Feluda | A configurable engine for analysing multi-lingual and multi-modal content, this repostory contains all the media processing algorithms relevant to facilitate search and clustering of media items for this project. |
Interested in contributing? Head over to CONTRIBUTING.md to learn more 🚦