AI Agents Enable Wallet Fraud
VeryAI highlights AI agents now transacting and managing wallets, creating risks for sybil attacks and spoofing — proposing Clawkey.ai with palm biometrics and zero-knowledge proofs for human verification. Identity Fusion warns of operational AI-driven threats including synthetic identities, deepfakes, and automated account takeovers requiring CIAM adaptation. Infocap notes persistent AI probes extending beyond initial onboarding.
The speed at which autonomous agents can move funds dramatically shrinks the window for intervention after a security breach. In 2025, illicit actors stole $2.87 billion through nearly 150 hacks, and in major incidents, the speed of post-compromise fund movement was a critical factor in preventing recovery. AI agents can automate the process of rapidly splitting and routing stolen funds across multiple blockchains, a task that once required significant manual effort. A Sybil attack, named after a case study of a woman with dissociative identity disorder, is when a single entity creates numerous fake identities to gain disproportionate influence over a network. In the context of AI, one person can deploy thousands of autonomous agents, making it nearly impossible for platforms to distinguish between genuine users and a malicious swarm designed to manipulate markets or drain funds. To counter this, identity verification is moving toward advanced biometrics and cryptography. Palm vein recognition, for instance, offers a false acceptance rate below 0.00001% by mapping millions of reference points in a person's hand. When paired with Zero-Knowledge Proofs (ZKPs), this allows an agent to prove it's operated by a real human without ever revealing the underlying private biometric data. The threat extends beyond wallets to AI-generated synthetic identities, which were linked to over $35 billion in fraud losses in 2023. The use of deepf
