AI and Cryptography: Experts Say the Threat Remains Unproven
AI could eventually challenge the mathematical assumptions behind encryption used by crypto wallets, banking and private messaging, but experts say current evidence does not show that today’s cryptography has been broken. In an a16z crypto discussion, Stanford cryptographer Dan Boneh, a16z crypto General Partner Eddy Lazzarin and Research Partner Justin Thaler examined what OpenAI’s recent mathematical research results do—and do not—demonstrate.
The panel said the potential risk from AI differs from the threat posed by quantum computing. Existing security practices and alternative cryptographic approaches offer reasons for confidence, they said, while systems should be designed to adapt if an underlying assumption fails. AI may also help researchers find flaws and developers build more secure software. For crypto markets, the discussion raises a long-term cybersecurity issue, not evidence of an immediate breach or a near-term threat to digital assets.
Neutral
The discussion presents no confirmed cryptographic break, exploit, or specific vulnerability affecting crypto wallets or blockchain networks. Its direct trading impact is therefore likely limited, with no clear basis for an immediate change in token valuations or market stability. Short-term reactions could include brief attention to cybersecurity narratives, but such attention is more likely to influence sentiment than fundamentals.
The longer-term issue is relevant: if AI advances eventually weaken cryptographic assumptions, developers and networks may need to adopt alternative algorithms and upgrade systems. That possibility could create migration costs and operational risk, while also encouraging investment in security research and cryptographic agility. The panel’s emphasis on uncertainty and existing safeguards, however, makes this a preparedness concern rather than a current market shock. As with past security debates around quantum computing, headlines may prompt discussion and volatility, but a credible technical demonstration or practical exploit would be needed to materially alter market behavior.