📽️
Digital Replicas and Talent ID
  • Digital Replicas & Talent ID: Provenance, Verification and New Automated Workflows
  • Overview
    • Executive Summary
    • The Challenge
    • A Digital Human is "Born"
    • Workflow
      • Asset & Metadata
      • Feature/Episodic Use Case
      • Usage Rights and Authentication Mapping
    • Legitimate Use Cases
    • Nefarious Use Cases
      • Deepfakes: By The Numbers
    • Legal & Ethical
    • Key Participants
      • Champions
        • ETC
        • ITV
        • Paramount Global
        • SMPTE
      • Contributors
        • Technology
          • EZDRM
          • The Scan Truck
          • Videntifier
          • WaveSeven
          • Wild Capture
        • Metadata & Governance
          • EIDR
          • HAND
          • C2PA
    • Additional Resources
Powered by GitBook
On this page
  1. Overview
  2. Nefarious Use Cases

Deepfakes: By The Numbers

PreviousNefarious Use CasesNextLegal & Ethical

Last updated 9 months ago

  1. 60% of consumers have encountered a deepfake video within the last year. Only 15% state that they have never encountered a deepfake video.

  2. Human detection of deepfake images averages 62% accuracy. Human subjects identified high-quality deepfake videos only 24.5% of the time.

  3. Only 38% of students have received guidance from their schools on how to identify AI-generated images, texts, or videos, despite a desire for such training expressed by 71% of students. A significant gap exists in educating students about identifying AI-generated content, with only 38% receiving guidance despite 71% expressing a need for such training.

  4. Only 22% of students feel very confident in their ability to detect whether an image they are viewing was generated with AI versus produced by a human. Many students lack confidence in discerning AI-generated images, highlighting a critical need for improved media literacy education.

  5. DeepFaceLab is used for over 95% of all deepfake videos. Available as open source code on GitHub, DeepFaceLab utilizes artificial neural networks to replicate visual and auditory features from an original video onto a target video. eftsure

Source: Jumio
Source: IEEE
Source:Center for Democracy & Technology
Source: Center for Democracy and Technology
Source: iperov
Deepfake statistics (2024): 25 new facts for CFOs - eftsure