Tackling Deepfakes in the new AI Legislative Framework
- Project team:
STOA (The Science Technology Options Assessment Panel of the European Parliament)
- Start date:
- End date:
- Project partners:
Rathenau Instituut (coordinator), Fraunhofer Institute for Systems and Innovation Research (ISI), Technology Centre CAS
- Research group:
The term “deepfake” is a combination of the terms “deep learning” and “fake” and refers to seemingly realistic media content (photo, audio, and video) which is altered and/or falsified by artificial intelligence techniques. The manipulation of media is not a new phenomenon in itself, but deepfakes use methods of machine learning and artificial neural networks to create fakes largely autonomously. Furthermore, easy-to-use and inexpensive software and apps support the spread of such products and services. In addition, the increasingly visual-based media consumption and the digital penetration of everyday life are leading to an intensification of this new phenomenon.
While there are certainly useful applications of deepfakes in the fields of education, art, and medicine, there are a number of risks, especially in the field of public opinion formation: These are not only on the individual level, but can also have an impact on society as a whole due to the potential disinformation and influence on democratic processes.
The project will conduct a literature review and expert interviews to examine deepfake techniques, future application scenarios, and their economic, social, ethical, and legal implications. Based on the results, it will then develop a set of policy options that are relevant to the committees of the European Parliament, including the new Special Committee on Artificial Intelligence in a Digital Age (AIDA), and relate to the current activities of the European Commission. In its White Paper, the European Commission has presented a framework for a future legislative proposal to tackle the risks posed by artificial intelligence.
Based on expert interviews with relevant stakeholders from academia, NGOs, and companies, the ITAS team will focus on the social, ethical, and legal aspects of deepfake technologies. It will analyze impacts on individuals, on important sectors and areas such as social media, but also consequences for society as a whole in terms of erosion of trust in information. In addition, ITAS will address the social and ethical dimensions of deepfakes and related policy options.
Karlsruhe Institute of Technology (KIT)
Institute for Technology Assessment and Systems Analysis (ITAS)
P.O. Box 3640
Tel.: +49 721 608-26133