AI Scams Outrun Blockchain Forensics as Losses Soar to $17 Billion

The development of blockchain forensics has improved significantly in recent years, allowing for better tracking and recovery of illicit funds. Tools like Chainalysis, TRM Labs, and Elliptic have frozen or recovered an estimated $34 billion in stolen money, with over 45 regulators worldwide using these tools as standard practice.

However, scammers have also become more sophisticated, particularly with the use of AI technology. According to Chainalysis, total crypto scam and fraud-related losses for 2025 sit at roughly $17 billion, up from $9.9 billion the previous year. What's more concerning is that AI-powered scams were found to be 4.5x more profitable than traditional ones.

AI scammers can now manufacture fake support agents, investors, or trusted insiders at scale, making impersonation fraud a growing concern. The FBI has warned of increasing impersonation scams, with Lior Aizik, co-founder and Chief Operating Officer at crypto exchange XBO, advising never to transfer funds to anyone who cannot be verified.

While defensive tooling has improved dramatically, the attack side has also become cheaper and faster. Scammers can now use AI to run targeted cons on profiled individuals, resulting in higher average payment sizes and increased profits.

The lack of safety layers integrated into AI trading agents is a significant concern, as seen with an incident where an agent bought a token that rugged 94% after twenty minutes, costing the wallet $12,000. The use of freeze authority, deployer history, and other red flags were not checked by the agent.

The damage caused by AI scams can be substantial, often without touching a smart contract at all. A woman lost $14,000 to scammers who impersonated YouTuber Mr Beast, highlighting the risk faced by public figures and their followers.