Phishing attacks in crypto are getting out of hand. Bitget Wallet just dropped new data showing phishing risk detections jumped 7.4x from Q2 to Q3 2026 — from 5K to 37K detections per week.

But here's the catch: they also rolled out AI-assisted monitoring in Q2, so it's unclear how much of this spike is real attacks vs better detection. COO Alvin Kan admits they can't separate the two.

What's clear: their AI is getting sharper. By Q4, it caught 90.2% of newly blocked risky domains — up from 36.3% in Q2. Nearly 70% of those domains weren't flagged by external threat intel databases.

Phishing-related signing risks made up 57.4% of all detections in Q3. Malicious addresses, risky contracts, and approvals followed.

Bitget's proprietary GetShield engine now simulates transactions, scans website code, and checks for hidden approvals before you sign anything. All free for users.

The reality: attackers are using AI to spin up convincing phishing sites at scale. Chainalysis says AI-linked scams generate 4.5x more revenue per operation.

No system catches everything, but if you're not using a wallet with real-time threat detection in 2026, you're playing with fire. Stay paranoid out there.