Digger – Deepfake Detection

Deepfakes are a real threat to society. Human inspection of audiovisual material is often not enough to make a judgement about its authenticity. With currently available technologies, it is not possible to reliably check whether the material has been altered.

The Digger project aims to use visual verification and audio forensics technologies to detect both shallow fakes as well as deepfakes or other types of synthetic media. The web-based verification platform Truly.Media, which was developed by Deutsche Welle and Athens Technology Center, will be extended with Fraunhofer IDMT’s audio forensics technologies. Using the platform, journalists will be able to carry out joint analyses to identify fake material. 

Responsibilities of Fraunhofer IDMT

As a technical  partner with experience in audio forensics, Fraunhofer IDMT will create and provide a secure REST API to be used for trustworthy audio analysis. The REST API will be implemented in the verification platform Truly.Media providing journalists access to the Fraunhofer IDMT audio forensics technologies for tampering detection. 

Partners

  • Deutsche Welle, Germany
  • Athens Technology Center, Greece

Sponsors

This project has received funding from the Google Digital News Initiative Fund (DNI Fund).

Duration

2019 – 2021