DeepFaceLab, DALL-E 2, Stable Diffusion and TikTok-filters are examples of the range of possibilities that artificial intelligence can currently achieve in the manipulation of images and videos - both positive and negative. For investigating authorities, public administration, the media, companies, and civil society, this also poses a wide range of challenges. Deep fake and image manipulation technologies are increasingly easily accessible and complex enough to be seen e.g. as a threat in cybercrime or as a challenge to democratic processes.
In the cooperative R&D project defame Fakes, we are therefore working on the detection of deep fakes and manipulation in digital image and video content and initiating awareness measures. The aim is to counteract the constant erosion of trust in digital content and to improve technologies for verification. The activities of the research project range from the generation of a data collection for the research of assessment tools, the evaluation of threat scenarios and the resulting requirements, to the iterative design of preventive and reactive measures. The focus is on the following three core areas:
- the development of suitable and comprehensible assessment tools for the detection ofdeepfakes and manipulations in huge collections of digital image and video data
- the conception and initiation of awareness measures in society to support the nationalimplementation of the "Deepfake Action Plan" with the close involvement of authorities, media,GSK partners and other relevant stakeholders and thus ultimately to take measures to raiseknowledge and awareness.
- analyzing the threat scenarios, the social, ethical, and legal implications of deep fakes and theassociated need for regulation.
Project Lead
Martin Boyer / AIT Austrian Institute of Technology GmbH
Partners
APA - Austria Presse Agentur eG
KSÖ Kompetenzzentrum sicheres Österreich
ÖIAT Österreichisches Institut für angewandte Telekommunikation
PwC Österreich
Bundesministerium für Inneres
Bundesministerium für Landesverteidigung
Contact
Martin Boyer / AIT Austrian Institute of Technology GmbH
Giefinggasse 4, 1210 Wien
Telefon: +43 50550 4284
Email: martin.boyer(at)ait.ac.at
Website: http://www.ait.ac.at/
Projektwebsite: www.defamefakes.at
