AI-Enabled Cryptocurrency Scams and Investor Vulnerability: Evidence on Scam Sophistication, Adoption, Financial Access and Losses

Abstract Cryptocurrency fraud is among the largest categories of financial crime, and generative AI may be raising the scale and persuasiveness of scams. We link AI-enabled scam sophistication, cryptocurrency adoption and financial access to investor losses, using public blockchain-analytics, survey and World Bank data. Scam operations linked to AI tooling earned about 4.5 times more than other operations, a descriptive comparison. Across eleven countries, loss prevalence ranges from 1% to 36%, and exposure alone does not explain the gap. Crypto adoption is unrelated to losses (Spearman ρ = 0.21; interval − 0.46 to 0.74). Account ownership is negatively associated with losses (ρ = −0.74), but is confounded with income and region. We discuss implications for regulators, exchanges and AI providers.