Changpeng Zhao, the founder of Binance, has cautioned that artificial intelligence may pose risks to Bitcoin users by exploiting leaked personal information and engaging in social engineering attacks, even without breaching the cryptocurrency’s cryptographic security. This information comes from reports by CoinMarketCap and Bitcoin Archive.
Summary
- CZ highlighted that compromised personal data can facilitate AI-driven phishing attempts without circumventing Bitcoin’s encryption methods.
- OWASP’s guidelines make comparisons of algorithms and key sizes, but do not verify claims regarding bank encryption security.
- Coinbase’s data breach in 2025 resulted in the exposure of user information while securing private keys and login details.
- Zscaler uncovered counterfeit Ledger verification pages designed to acquire users’ wallet recovery phrases.
According to CoinMarketCap, Zhao’s warning emphasizes the need for awareness about more sophisticated phishing tactics made possible by the availability of personal information. Bitcoin Archive also reported that he mentioned banks employing weaker encryption techniques compared to Bitcoin. However, neither of these sources provided details on the algorithms, key sizes, or banking systems necessary for validating this claim.
FBI Issues Guidance on AI-Enhanced Financial Frauds
The FBI’s December 2024 advisory outlines how criminals leverage generative AI to enhance the credibility of fraud schemes, enabling them to reach potential victims more effectively. Examples range from fake cryptocurrency investment platforms to impersonated social media accounts and chatbots directing users to harmful links.
According to the agency, offenders can utilize generated content and translations to eliminate mistakes in spelling or grammar that might otherwise reveal a scam. The guidance also highlights the use of synthetic images for creating false identities and counterfeit identification documents, as well as initiating video calls to impersonate high-ranking individuals.
The FBI cautioned that cloned audio could mimic family members requesting emergency funds or impersonate bank clients. These warnings focus on deception and identity theft rather than suggesting that AI has breached Bitcoin’s cryptographic safeguards.
For individuals, the agency recommends establishing a unique code or phrase to verify a relative’s identity. It also advises against publicly sharing personal images and audio recordings that could be exploited to create convincing impersonations.
Evaluating Bitcoin Security Claims Requires Specific Evidence
OWASP’s Application Security Verification Standard provides an analysis of cryptographic strength across algorithms and key sizes, employing a comparison table derived from guidelines by the US National Institute of Standards and Technology. This table categorizes AES-128, 3,072-bit RSA, and elliptic-curve cryptography using keys ranging from 256 to 383 bits into the same 128-bit security bracket.
OWASP’s framework does not position banks and Bitcoin as equivalent systems. Instead, it urges organizations to identify and document their cryptographic assets—including algorithms, keys, and certificates—and evaluate them as cryptographic practices and needs evolve.
The comparison of banks to Bitcoin, as attributed to Zhao, lacks concrete evidence at the system level. The two posts do not specify a bank, denote a cryptographic vulnerability, or confirm an AI-driven attack that accessed a Bitcoin private key.
Quantum computing raises a different technological issue in NIST’s published directives. In August 2024, the agency released three post-quantum standards aimed at safeguarding electronic data against assaults from sufficiently advanced quantum computers. The announcement describes systems that operate distinctly from traditional computers and could potentially dismantle commonly used cryptographic techniques.
NIST indicated that these standards encompass general encryption and digital signatures, applicable in scenarios ranging from secure communications to online transactions. The agency urged system administrators to begin implementing these standards, as the integration process may require time.
Coinbase Shares Details on Stolen Data Used for Impersonation
A US cryptocurrency exchange has reported on a customer-data breach utilizing a similar methodology. In May 2025, Coinbase mentioned that perpetrators had bribed support personnel abroad to access information that facilitated impersonation of the company, as previously detailed in its report on the data breach.
Coinbase’s disclosure identified contact information, images of governmental ID, masked financial identifiers, and account details among the breached records. The company emphasized that the attackers did not gain access to private keys or login credentials, and that the compromised data alone did not allow for the movement of customer funds.
The exchange rejected a ransom demand and publicly announced a $20 million reward for information leading to the arrest and conviction of the attackers. The disclosure included the implementation of measures like additional identity verifications for large withdrawals and fraud warnings for suspicious accounts.
Coinbase reported that it would offer voluntary reimbursements to retail customers who sent funds to the scammer as a result of the incident, pending review. However, the disclosure did not confirm AI’s involvement in these attacks.
Phishing Pages Target Ledger Users for Wallet Recovery Information
Users of hardware wallets have encountered impersonation tactics via search engine advertisements. A Zscaler investigation on September 25, referenced in an October 11 article regarding fraudulent Ledger websites, revealed deceptive Google ads directing users to a counterfeit wallet interface.
Zscaler documented a phishing campaign analyzed in August, involving redirects through Google Cloud Storage and Vercel before presenting a fake verification page hosted on Google Sites. Researchers noted that these redirect URLs switched approximately every 15 to 20 minutes.
The fraudulent verification process prompted users to enter their recovery phrases, allowing attackers to access wallets without the need for physical devices. Zscaler’s report detailed credential theft but did not establish AI’s role in this operation.
Ledger has issued its own warnings regarding counterfeit communications, which may include physical letters prompting users to scan QR codes or navigate to websites before entering their 24 recovery words. The company cautions that possession of this phrase grants access to any associated accounts and stresses that it should never be entered on a computer.
The company’s safety guidelines state that it does not send out emails urging product firmware updates, setting up two-factor authentication, or taking similar actions. Its warning page also includes illustrations of fake letters and emails that attempt to solicit recovery information.