arstechnica.com 7/31/2026, 2:20:39 PM · external

AI chatbots beat humans in pig butchering scams, study shows

AI chatbots beat humans in pig butchering scams, study shows
CyberSIXT Evidence Panel Source marked as original reporting

A study by researchers from four universities indicates that AI chatbots can outperform human scammers in building trust during long con scams, specifically in the "pig butchering" fraud scheme that scams victims out of significant sums through fake investments. In a simulated environment, the AI chatbots were more effective at persuading subjects to engage and download a test app, achieving a 46% compliance rate compared to 18% for human scammers.

The AI was graded with higher trust scores by test subjects and managed to convincingly impersonate humans without revealing its identity. This suggests that AI could soon automate much of the scamming process, potentially reducing the need for human trafficked workers in scam operations, but could also complicate efforts to track and combat these schemes.

View full article

Article by CyberSIXT