Artificial Intelligence for Deepfakes Creation and Detection: Friends or foes?

December 2021 | 42 pages | ID: A1B021E31762EN
IDATE

Abstracts

This report is an all-encompassing analysis of the current state of the art of deepfakes and their underlying technology, the contexts in which deepfakes are and can be used, the impact of deepfakes in the various areas of application, as well as active initiatives and cutting-edge techniques used to detect them.

It focuses on artificial intelligence and its major subfields, identifies deepfakes creation and detection examples including use cases and provides a quick overview of the national initiatives developed to tackle deepfakes.

Key questions answered in the report:

• What are “Deepfakes” and what is the technology behind them?
• What are their areas of application?
• What techniques can spot them and who are the major players engaged in developing such techniques?
• What are countries doing to face the threats deepfakes bring to the societies?

Contents

1. KEY FINDINGS

2. WHAT DEEP LEARNING HAS CHANGED

2.1. Introduction of concept
  Artificial Intelligence
  Machine Learning, a branch of Artificial Intelligence
  Deep Learning, a branch of Machine Learning
  Deep Learning, an AI game changer
2.2. Impact of deep learning

3. EXPLAINING DEEPFAKES

3.1. Deepfakes
3.2. Implications of deepfakes
3.3. Deepfakes areas of applications
3.4. Positive and neutral impacts of deepfakes
3.5. Deepfakes software
3.6. Conception of deepfakes
3.7. Deepfakes creation offering

4. AI, GENERATING DISINFORMATION

4.1. Benchmark of deepfakes creation
4.2. Deepfakes web – face-swap
4.3. REFACE – face-swap
4.4. Modulate.ai – voice cloning
4.5. Rephrase.ai – deepfake videos speaking any new text

5. AI, DETECTING DEEPFAKES

5.1. Major national initiatives that tackle misinformation
5.2. The common techniques of deepfake detections
5.3. Deepfakes detection techniques
  Facebook’s AI tool
  Microsoft Video Authenticator
  InVID & WeVerify
  Project Origin
  The News Provenance Project
5.4. Benchmarking of active initiatives on detecting deepfakes and disinformation