Binance: AI prevented $10.53B in crypto scam losses

Binance reported that its artificial intelligence security tools prevented $10.53 billion in potential user losses from crypto scams and phishing between Q1 and Q2 2025, according to a company blog that detailed roughly two dozen AI-powered security features.

The blog described tools that combine computer vision and language analysis to spot fraud. “Computer vision is used to detect fake payment proofs, while real-time language analysis helps surface scam patterns in P2P transactions,” the post said. It also noted that automated decisioning now covers 57% of the exchange’s fraud controls and that those controls produced a 60–70% reduction in card fraud rates compared with industry benchmarks. The company added that its identity verification systems are being updated to detect deepfakes and synthetic identities, with reported efficiency gains of up to 100x over manual processes without AI.

Binance included more recent operational figures for 2026. In Q1 2026 the exchange reported protecting $1.98 billion from 22.9 million scam and phishing attempts, recovering $12.8 million across about 48,000 cases, confiscating $131 million in illicit funds and processing more than 71,000 formal requests from law enforcement.

The blog outlined coordination with other firms through a joint security unit known as T3. The unit froze $344 million in USDT that Binance linked to Iranian entities. The exchange also described a new withdrawal lockdown feature designed to reduce the risk of wrench attacks, where criminals coerce victims to authorize transfers.

Binance Research, cited in the blog, estimated that AI is currently about twice as effective at exploitation as at detection and that AI-enabled scams are roughly 4.5 times more profitable than traditional scams. The post referenced industry examples, including an estimate from a major bank that its AI systems helped prevent about $1.5 billion in fraud losses, and a blockchain security firm’s finding that crypto-motivated physical assaults are on pace to exceed the 2025 total.

The blog addressed recent scrutiny of Binance’s surveillance and compliance systems following reports that employees who flagged transfers to Iran-linked entities were dismissed. The exchange denied the allegations and said it cooperates with U.S. regulatory and compliance agencies.

The company’s public post listed the AI features, case numbers and enforcement actions and described how automated systems are used to triage threats and support law enforcement requests.