CZ Downplays Risk of AI Breaking Bitcoin, Warns of Phishing and Data Leaks

Changpeng Zhao of Binance has stated that cracking Bitcoin’s encryption is not required for AI to inflict harm. Leaked data, in his view, already empowers phishing and social engineering schemes.

Bitcoin Core maintainers, meanwhile, have altered their security disclosure practices. They cite the increased ease with which AI can identify vulnerabilities as the reason for the shift.

Everyone worried about AI hacking crypto, but the internet and banking systems rely on weaker encryption.

That said, AI doesn’t need to break Bitcoin’s cryptography to cause damage, leaked data already enables more phishing, social engineering, etc,

Stay SAFU!

— CZ 🔊 BNB (@cz_binance) October 23, 2026

Phishing a Bigger Concern Than Code Breaking, CZ Argues

In an X post on October 23, Zhao (commonly referred to as CZ) put forward his argument. He contends that the weakness does not lie in cryptocurrency encryption.

Furthermore, CZ indicated that banks and much of the internet depend on encryption that is weaker than Bitcoin’s. Consequently, those legacy systems would be at greater risk should AI code-breaking become a wider problem.

The ex-Binance boss noted that AI currently enhances the credibility of fraud attempts. Malicious actors pair previously exposed personal information with sophisticated AI-generated correspondence appearing to originate from known individuals. Zhao concluded his message with a regular admonition: “Stay SAFU!”

CZs perspective diverges from that of Ethereum researcher Justin Drake, who on October 8 called on the sector to “calmly start planning for ‘bunker mode.’” Drake posited a scenario where AI could compromise the mathematical protections of wallets “in months, not years.”

No functional exploit has materialized yet, nonetheless. Drake’s advice was for major asset holders to progressively transfer wealth to wallets whose public keys have never been broadcast on the ledger.

Today I call upon the blockchain industry to calmly begin planning for "bunker mode". My personal recommendation is to set in motion a controlled mass migration of assets to fresh addresses, i.e. addresses whose pubkeys remain hidden behind a hash.

Holders, starting with large…

— Justin Drake (@drakefjustin) October 7, 2026

Bitcoin Core Evolving Strategy for AI-Detected Flaws

Developer Antoine Poinsot unveiled a revised security strategy on the Bitcoin Development Mailing List on Thursday. He explained that large language models render certain vulnerabilities trivial to find again.

Consequently, Poinsot argued that lengthy confidentiality periods for such bugs are mostly ineffective, and merely postpone the distribution of patched versions to the user base.

The updated protocol dictates that readily findable weaknesses will be rectified promptly across all supported editions. Security notifications will accompany these corrections.

Other categories of defects adhere to schedules determined by their gravity. Low-importance problems are disclosed a fortnight after a major upgrade incorporates the solution. Medium- and high-severity errors remain confidential until the final impacted version reaches the end of its life. Critical vulnerabilities are anticipated to require a distinct handling procedure.

Collectively, these shifts illustrate the rapid pace at which artificial intelligence is transforming security within the cryptocurrency sphere, spanning user-level fraud to foundational software.