In the year-end report for its Anti-Scam Refund Initiative, Binance shared how it has applied cutting-edge artificial intelligence (AI) and machine learning (ML) technologies to transform digital asset security from a reactive measure to a proactive defense strategy.

Anti-Scam Initiative

According to the report, the company processed an average of 80 successful refunds per month, totaling approximately $9 million in stolen funds returned to victims in 2024. Additionally, the company reported making over 30,000 phone calls to warn potential victims of possible scams and sending at least 15,000 alerts to platform users daily.

The initiative resulted in the blacklisting of at least 47,000 malicious addresses and the prevention of over $129 million in fraud per year as of November.

Key to Binance’s approach to preventing bad actors from stealing from its user base is a system that combines technological surveillance and human-centered support. It uses machine learning algorithms to analyze complex transaction patterns in real-time, identifying potential criminal activity at lightning speeds. It also uses AI-powered behavioral profiling to differentiate between legitimate user activity and potentially illegal activity.

The company said it has developed over 50 specialized models and implemented 14 major enhancements to outsmart increasingly sophisticated fraudsters.

Its anti-fraud fund recovery initiative operates on four pillars: proactive protection, 24/7 security mechanisms, rapid response recovery, and support for silent victims. A one-day safety net provides users with a cooling-off period for suspicious transactions, with funds moved to flagged accounts and frozen to allow for investigations and potential intervention.