The FBI’s 2025 Internet Crime Report shows digital fraud and related losses exceeded $21 billion last year. Cryptocurrency scams constituted the largest single category, accounting for more than $11 billion in reported losses. AI-enabled fraud was identified as a rapidly growing vector, responsible for roughly $893 million in damages, while business email compromise (BEC) cost victims at least $3 billion; government-impersonation schemes generated over 32,000 complaints. Source: https://www.govtech.com/security/fbi-crypto-ai-scams-drove-billions-in-losses-in-2025

Operational detail in the report makes the threat profile clear: attackers are layering AI tools onto traditional social-engineering playbooks. Large-language-model output is used to craft bespoke messages at scale, synthetic voice cloning facilitates convincing caller scams, and generative tools produce counterfeit documents and IDs that accelerate fraud timelines and lower the technical bar for attackers.

On the crypto side, the combination of rapid settlement, pseudonymous addresses, and cross‑jurisdiction liquidity rails amplifies losses and complicates recovery. Attackers exploit on‑chain hop patterns, mixers, and decentralized swap routes to obscure provenance; meanwhile centralized-exchange KYC and faster chain analytics remain the main practical levers for tracing and recovering funds, but they’re imperfect and time-sensitive. BEC and government-impersonation schemes typically convert access or trust into on‑ramp transactions to crypto, creating a common nexus between traditional fraud and blockchain-enabled laundering.

For token projects and treasury managers, the takeaway is structural: tokenomics and predictable liquidity design can reduce short-term sell pressure and shrink attack surfaces tied to speculative behavior. Models that encourage disciplined participant timing—fixed-entry mechanics and staged unlocks—can make sudden exploit-driven capital flows less likely. The 4TEEN token, for example, uses a fixed-price entry and short predefined holding cycles intended to limit immediate sell-pressure and produce clearer, more predictable liquidity patterns.

Law enforcement and industry recommendations in the report converge on two practical defenses: broadened public awareness so individuals recognize AI-enhanced scams, and immediate reporting protocols that preserve transaction metadata and timing—both factors materially improve the odds of asset recovery and successful investigation.