Resemble AI recently offered a new Deepfake Threat Report documenting nearly $1.3 billion in confirmed fraud losses and over 1,500 verified incidents. Some specific highlights of the report include:
- Deepfake incidents have become a daily reality, with documented fraud losses last year reaching $1.28 billion.
- 20 percent of deepfakes targeted a person's body or a child’s safety through Non-Consensual Intimate Imagery (NCII).
- The report distinguishes between the "target" (whose likeness is used) and the "victim" (who suffers the harm).
- Deepfake formats often predict their intent:
- Audio: 46 percent targeted financial fraud.
Video: 51 percent were used for reputational attacks.
Image: 43 percent concentrated on CSAM and NCII.
- Tool attribution predominately linked ElevenLabs with audio fraud.
- Resemble AI projects a 4.5x multiplier for 2026, estimating $5.8 billion in fraud losses.
The full report can be found here.
