Binance: AI security blocked $10.53B in crypto scams

Binance reported that its artificial intelligence security tools blocked $10.53 billion in crypto scams and fraud between the first quarter of 2025 and March 2026 and blacklisted 36,000 malicious addresses. The exchange said more than 5.4 million users were protected during the period after it rolled out over 24 AI-driven initiatives and more than 100 machine-learning models.

The company framed the figures as covering a 15-month period ending in March 2026. In the first quarter of 2026 alone, the exchange intercepted 22.9 million scam and phishing attempts and estimated it prevented $1.98 billion in user losses during that quarter.

Binance described specific technical defenses in use. Computer vision was applied to detect forged payment proofs, real-time language analysis was used to identify common scam patterns, and identity verification tools were enhanced to counter deepfakes and synthetic identities. The exchange reported that AI-driven decisioning now powers 57% of its fraud controls and that those measures corresponded with a 60% to 70% reduction in card fraud rates compared with industry benchmarks.

The company warned that lower technical barriers and advances in AI are enabling more automated and scalable social-engineering attacks. In a blog post, Binance wrote: “AI-powered scams and exploits are accelerating. The barrier to entry for scam perpetrators is falling fast, with AI accelerating the drop. What once required technical expertise can now be executed for next to nothing and at scale.” The post added that “AI is amplifying social engineering at an unprecedented level, powering deepfakes, phishing bots, fake platforms, voice cloning and impersonation across chat applications, exploiting trust and urgency.”

The report referenced broader fraud figures for context. The FBI reported in April that U.S. citizens lost about $11 billion in crypto to scams, with impersonation of officials and firms a common tactic.

Binance reported its systems not only blocked transactions and accounts tied to fraudulent activity but also integrated AI into monitoring and response workflows to identify and blacklist suspicious addresses. The exchange presented the data as an account of attempted losses intercepted by its technology and internal controls during the 15-month period.