AI-Powered Crypto Scams Are Earning 4.5x More Than Traditional Schemes Here's How They Work
-

Crypto has become one of the primary financial targets for AI-driven deepfake and impersonation scams, and the economic data explains why fraudsters are investing in the technology. According to Chainalysis, the average AI-assisted crypto scam nets approximately $3.2 million — roughly 4.5 times the yield of a conventional scheme that does not use AI tools. That return on investment makes the cost of consumer-grade deepfake software trivially small relative to the potential payout, creating a strong and growing financial incentive for criminal operations to integrate AI into their playbooks. The mechanics are increasingly sophisticated: fraudsters pair deepfakes and face-swap applications with large language models to enhance classic romance and investment scams, using AI-generated video of real people to establish false trust before executing the financial deception. In August 2025, attackers stole $2 million by impersonating the founder of Plasma using deepfake technology. North Korean operatives have been documented running deepfake video calls on Zoom as part of crypto-targeted social engineering operations.
The threat is particularly acute in crypto because the sector's culture of digital-first interaction, pseudonymity, and irreversible transactions creates an environment where deepfake impersonation is both easier to execute and harder to recover from than in traditional finance. When a victim sends crypto to a scammer impersonating a project founder or investment platform, there is no fraud reversal mechanism, no chargeback process, and often no regulatory body with jurisdiction to pursue recovery. The combination of consumer-grade deepfake tools that can convincingly fake real-time video calls, large language models that can sustain coherent long-form deceptive conversations, and crypto's structural irreversibility creates a threat profile that is genuinely difficult to defend against at the individual level. The most effective protection remains behavioral rather than technical: treating any unsolicited investment opportunity with extreme skepticism regardless of how convincing the video or voice communication appears, independently verifying the identity of anyone requesting crypto transfers through out-of-band channels, and recognizing that the visual and audio authenticity of a video call is no longer a reliable signal of the person's actual identity.