The Deepfake Threat Map
Deepfakes are attacks that deceive people with AI-generated, real-looking video, audio, and images. Until 2024 the "spread widely to influence" type, like fake politician videos, was central, but from 2025 scams pinpointing individuals and companies have surged.
- Money-transfer-instruction calls imitating a CEO's voice (damage cases of hundreds of millions of yen)
- Fraud using deepfake video in hiring interviews so a different person continues to work
- The trick of using a family member's voice to make someone wire "settlement money for an accident"
- Generating sexual images without consent (turned into persistent blackmail material)
