AI Scams Target Crypto Users, Not Protocols, and Wallet Security Is Catching Up
AI-powered scams targeting crypto users surged in 2025, leading to billions in losses.
Run Chainalysis?
Get an email when a reviewed story names Chainalysis, usually within the hour.
Free. Your list stays private and never appears in a subject line. One click stops it. How Vendor Watch worksPrivacy
Summary
AI is enabling a significant increase in crypto scams, with impersonation schemes growing over 1,400% in 2025. These scams leverage AI-generated voices and faces to create convincing lures that exploit user emotions like fear and urgency, rather than technical vulnerabilities. Despite a decrease in phishing incidents, the total losses remain high, indicating a shift towards fewer, more impactful attacks.
The lesson from this incident
Full lessonAI tools are dramatically lowering the barrier for cybercriminals to craft highly convincing impersonation scams targeting cryptocurrency users, with incidents rising over 1,400% in 2025. Rather than exploiting technical flaws in protocols, attackers weaponize synthetic voices, deepfake faces, and emotionally manipulative messaging to trick users into surrendering funds or credentials. The shift toward fewer but higher-impact attacks suggests scammers are investing more effort per target, making the threat harder to detect with volume-based defenses. This underscores that the human element remains the most exploitable attack surface in the crypto ecosystem, and that user education and strong authentication controls are now frontline defenses.
What helps against this
How we list vendors- Phishing-resistant MFA
A stolen password, or a user who taps Approve on a push prompt, is no longer enough to log in.
- Email security
Phishing, malware and invoice fraud are filtered out before they reach inboxes.
- Security awareness training and phishing simulation
People recognise and report attacks instead of falling for them.
Indicators of Compromise
- malware — face-swap software
- malware — deepfake tools
- malware — large language models