Resemble AI · Annual Threat Intelligence
Now Available
A breakdown of global incidents, their impact and the implications of the growing AI deepfake crisis.
- 1,567 unique verified deepfake incidents — generating 296.4B combined media impressions across 3,253 reported incidents
- 80%+ of deepfake incidents did not disclose financial damage — so while $1.28B in losses were reported, the true financial toll is likely far greater
- 20% of all incidents involved CSAM or NCII — 311 unique incidents targeting real people’s bodies or children’s safety
- Zero lag between a major world event and the first deepfake response — deployment within hours became the norm in 2025
- 350× the reach of the top 5% of Critical incidents over Minimal incidents — generating 68% of all media reach. A brand targeted at this tier averages 2.8B impressions per incident
Section 01: The attacks with the highest reach are Brand & Reputation
In 2025, deepfake incidents collectively reached an audience of 296.4 billion — roughly 37 times the global internet population. That reach is sharply concentrated: 54 unique incidents with five or more sources averaged 1.5 billion impressions each, while 81% of all unique incidents were covered by a single outlet.
- 296.4B total media reach across 1,567 unique incidents (~3,253 reported)
- Brand & Reputation attacks generated 156.3B in reach — more than all other categories combined
- 81% of unique incidents were single-source, averaging 112M reach each
- 54 unique incidents with 5+ sources averaged 1.5B reach — 13.4× the single-source average
- Fraud incidents caused the majority of documented financial losses while generating a fraction of Brand & Reputation reach
Media Reach by Attack Type
Combined reported reach by attack type. Brand & Reputation attacks, which include fake endorsements, impersonation and defamation, dominate reach.
- Brand & Reputation: 156.3B
- Political Disinformation: 41.2B
- NCII: 37.9B
- Consumer Fraud: 26.5B
- CSAM: 22.1B
- Corporate Fraud: 8.5B
While the reach of corporate fraud looks small by comparison, these attacks often happen by defrauding employees and are only surfaced by an inside source or whistleblower. Despite attempts to keep them quiet, corporate fraud attacks still reached nearly 9 billion global impressions throughout the year.
Section 02: $1.28 billion in fraud losses documented
In 2025, deepfake fraud cost documented victims over $1.28 billion. Corporate and consumer fraud operate at different scales, hit different victims, and leave very different footprints.
- Corporate Fraud: 41 unique documented incidents; $74.9M documented losses; largest single incident $35M; 71% with no reported damage
- Consumer Fraud: 289 unique documented incidents; $1.34B documented losses; largest single incident $700M; 83% with no reported damage
Continue reading — 10 more pages of data
CSAM & NCII trends, political disinformation, target profiling, platform response, and the complete methodology. Delivered to your inbox instantly.
1,567 verified incidents. Deduplicated & peer-reviewed. Free — no credit card.
