The FBI’s 2025 Internet Crime Report shows U.S. losses from cyber-enabled crimes approaching $21 billion, with the Internet Crime Complaint Center (IC3) receiving 1,008,597 complaints — a year-over-year increase that underscores both scale and acceleration in digital fraud. Source reporting on the FBI release is available here: https://www.hstoday.us/subject-matter-areas/cybersecurity/cryptocurrency-and-ai-scams-cost-u-s-almost-21-billion-in-2025-according-to-new-fbi-internet-crime-report/

What the numbers say

• Total complaints: 1,008,597 (IC3).
• Aggregate losses: nearly $21 billion in 2025.
• Cyber-enabled fraud: losses exceeded $17.7 billion.
• AI-related complaints: 22,364, resulting in nearly $893 million in losses.

Operational response and outcomes
To tackle crypto-related fraud, the FBI launched Operation Level Up, an enforcement and victim-notification effort that alerted more than 8,000 potential victims and helped recover or prevent losses in excess of $500 million. In 2026 the Bureau introduced Operation Winter SHIELD to expand digital-security initiatives and cross-jurisdictional coordination, signaling an intensification of law-enforcement focus on novel tech-enabled crimes.

Tactics scaling with technology
The report documents a clear shift in attacker tradecraft. Scammers increasingly combine traditional social-engineering pressure with AI-augmented tools:

• Fake social profiles and elaborate persona farms to establish trust.
• Voice cloning and synthetic video to impersonate trusted contacts or public figures.
• Highly convincing deepfakes used to authorize transactions or coerce wire/crypto transfers.
• Time-pressure and escalation tactics to short-circuit victims’ due diligence and force rapid asset transfers.

Market and protocol mechanics at risk
Cryptocurrency rails remain an efficient conduit for fraud because of speed, pseudonymity, and the multiplicity of custody choices. Scams exploit on-ramps/off-ramps, peer-to-peer liquidity, and poorly designed token incentives — especially where early participants can extract value quickly. Poor tokenomics and open liquidity channels amplify pump-and-dump dynamics: when selling pressure concentrates and exits are friction-free, it’s easier for bad actors to realize gains and for victims to absorb losses.

A design contrast
Token structures that limit immediate sell-pressure and create predictable liquidity behavior reduce some exploit vectors. For example, certain fixed-price entry models with short, defined holding cycles are intended to lower instantaneous sell-pressure and align participant timing, which can make coordinated exits and rapid rug pulls harder to execute.

Implications for practitioners

• Custodial choices matter: stronger custody and withdrawal controls, combined with forensic-ready transaction logging, reduce recoverability friction and deter abuse.
• Tech literacy and vendor controls: organizations and retail users need clearer guidance on verifying identity claims and detecting synthetic media.
• Protocol-level defenses: tokenomic safeguards that temper impulsive liquidity events and reward disciplined participation can change exploit economics so scams are less profitable.

AI’s role is dual-use
The AI category generated 22,364 complaints and nearly $893 million in losses, illustrating how generative tools lower the cost and raise the fidelity of impersonation. This makes remediation and attribution harder for investigators while increasing the speed and scale at which social-engineering campaigns can run.