FBI: 22,364 AI-fraud complaints, $893M in losses during 2025
The FBI Internet Crime Complaint Center (IC3) documented 22,364 AI-related complaints across the United States in 2025, with adjusted losses totaling $893,346,472, FBI Internet Crime Report 2025. The report, released in April 2026, marked the first systematic assessment of AI-enabled fraud at scale, showing that autonomous and AI-assisted attack methods have become a measurable vector in U.S. cybercrime.
Scale of AI-Powered Crime
The 22,364 complaints represent a discrete category within IC3's broader fraud tracking. Each complaint documents a victim report of financial harm or attempted fraud involving artificial intelligence—whether deepfakes used in impersonation schemes, voice cloning for social engineering, or autonomous systems deployed to scale phishing campaigns. At an average loss per complaint exceeding $39,900, the aggregate $893 million figure reflects the material impact on individuals, businesses, and institutions.
Social Engineering Examples and InfoSec Federation reporting on the FBI data both highlighted that AI-generated content—particularly synthetic media—has enabled attackers to scale social engineering attacks beyond what was operationally feasible in the pre-agent era. The FBI's dedicated section in the report titled "Artificial Intelligence (AI) Used in Cybercrime" signals institutional recognition that AI capability is now a distinct criminal tool category.
Verification and Scope
The figures are grounded in IC3's complaint intake system, which receives reports from U.S. victims of internet crime. The adjusted-loss figure accounts for validated victim reports and excludes unverified claims, making it a conservative measure of actual fraud. No specific court sentences or regulatory actions were announced alongside the report—the release was statistical, not enforcement-driven. However, the publication of these baseline metrics establishes a data foundation for law enforcement and policy makers tracking AI-enabled fraud trends.
Real-World Agent Actions
While the FBI report does not name individual cases, the scale of complaints confirms that AI agents and autonomous systems are now operationalized in fraud pipelines. The 22,364 complaints span deepfake video and audio impersonation, automated credential-stuffing attacks, AI-generated phishing content, and identity-theft schemes assisted by synthetic identity generation. These are not hypothetical capabilities; they are documented incidents reported by victims and compiled by federal law enforcement.
What's Next
The IC3 report serves as a baseline for tracking whether AI-enabled fraud complaints and losses grow, stabilize, or decline in 2026 and beyond. It also informs threat prioritization at the FBI, Department of Justice, and state attorneys general offices investigating organized cybercrime networks. For enterprises and individuals, the published figures reinforce that AI-assisted fraud is a material risk vector, not a future concern.