How Social Engineering and AI Can Enable Crypto Attacks

Artificial intelligence (AI) is making cryptocurrency fraud more convincing, while social engineering continues to exploit the weakest link in digital asset security: human trust. Criminals increasingly impersonate executives, customer support representatives and investment advisers to steal credentials, manipulate wallet approvals and redirect funds. Recent industry reports suggest attacks are becoming more targeted rather than simply more frequent.

Crypto Fraud Losses Reveal Growing Threat

Blockchain analytics firm Chainalysis estimated that cryptocurrency scams generated as much as USD 17 billion in 2025. Impersonation scams increased more than 1,400% year over year, while AI-enabled operations generated 4.5 times more revenue than scams without identified AI involvement.

The average scam payment reached USD 2,764, compared with USD 782 in 2024, representing a 253% increase. This indicates that convincing victims to authorize transactions can be highly profitable without breaking blockchain protocols.

How AI Strengthens Social Engineering

Attackers can use generative AI to write personalized phishing messages, create synthetic identities and imitate familiar voices during calls. Deepfake videos can make fraudulent investment endorsements appear credible, while fake exchange websites capture login details and recovery phrases.

A common attack begins with an urgent message claiming an account is compromised. The victim is then directed to a counterfeit support portal or persuaded to approve a malicious wallet transaction. As blockchain transfers are generally irreversible, recovery can be difficult once assets leave the wallet.

Wallet Compromises and Phishing Dominate

According to CertiK's first-half 2026 security report , Web3 incidents caused USD 1.31 billion in losses across 344 cases. Wallet compromises accounted for approximately USD 444.5 million across 33 incidents, while phishing generated USD 366.3 million across 63 incidents.

Phishing incidents declined 52.3% compared with the year's first half, but associated losses dropped only 10.8%. Four attacks accounted for approximately 85% of phishing losses, highlighting the financial impact of attacks on carefully selected targets.

Regulators Warn Automated Attacks

On November 4, India's CERT-In warned that an AI-assisted campaign dubbed CRAFTY FALCON was targeting financial institutions, including cryptocurrency service providers. Its advisory described stolen wallet keys, unauthorized transfers and automated exploitation of payment systems.

The warning underscores that AI can support both deception and technical intrusion, creating risks for individuals and exchanges alike.

Protecting Crypto Assets

Investors should independently verify support requests, avoid sharing seed phrases and scrutinize wallet permissions before signing. Hardware wallets , phishing-resistant authentication and transaction limits can reduce exposure.

Exchanges should strengthen withdrawal approvals, employee verification and suspicious-activity monitoring. Security teams should also rehearse deepfake scenarios and require independent confirmation for unusual fund movements. As AI lowers the cost of convincing impersonation, layered security and deliberate verification remain essential defenses against increasingly sophisticated cryptocurrency attacks.

Also Read: Binance Hack: Elliptic Traces USD 124M in Bitcoin Swaps Across 16 Blockchains