Avalanche founder Emin Gün Sirer warned that artificial intelligence could help attackers uncover previously unknown weaknesses in blockchain code, software libraries and wallets before existing cryptographic systems are broken. The concern shifts attention from the future failure of cryptography to the security of today’s software. Sirer did not identify an unresolved vulnerability in the XRP Ledger or present evidence of an AI-driven attack, making his comments a risk warning rather than a disclosed security incident.
Software flaws may emerge before cryptography fails
In a post on X, Sirer said AI was more likely to be used first to find software-level weaknesses than to render widely used cryptographic systems such as ECDSA obsolete. More capable models could identify problems missed during human review, including flaws in blockchain software, third-party dependencies and user wallets.
If attackers find those weaknesses first, they could exploit them before development teams complete a fix. Sirer said such vulnerabilities could potentially lead to losses of tens of billions of dollars, potentially before AI can break widely used cryptographic methods. The scenario remains a description of a possible risk: the information available does not identify a specific vulnerability, affected software version or confirmed loss.
XRP Ledger faces scrutiny, but no AI attack has been disclosed
Sirer’s warning came as the XRP Ledger again drew security attention following an emergency software update. However, available information does not establish a direct link between the update and an AI-discovered vulnerability, nor does it disclose the specific issue addressed by the update. The update therefore does not establish that the XRP Ledger was attacked or that the network has a confirmed unpatched flaw.
The impact of AI-assisted vulnerability discovery would depend not only on whether a model can find a defect, but also on whether the defect can be exploited, how quickly the project team issues a fix and whether users update their software. For people holding or using on-chain assets, the underlying code, wallets and dependency components can all affect security. Sirer did not provide verifiable details of an XRP Ledger vulnerability in his comments.
Crypto leaders debate AI’s impact on security
Ethereum co-founder Vitalik Buterin has also warned that mathematical discoveries accelerated by AI could pose a threat to cryptography. Sirer’s argument points to a different risk path: even if core cryptographic systems remain intact, weaknesses in software implementations could become an attack route first. The two concerns involve different parts of the technology stack and should not be treated as equivalent.
The discussion follows earlier public exchanges among crypto industry figures. Ripple CEO Brad Garlinghouse previously criticized Sirer’s comments about Ripple’s progress in banking. That dispute concerned banking activities rather than security, but the latest warning shows that public discussion of Ripple and its ecosystem among industry leaders remains active.
Based on the information currently available, Sirer’s comments amount to a warning about software-security risks in the age of AI. He described how attackers might exploit undiscovered defects, but did not publish evidence of a vulnerability in the XRP Ledger or demonstrate an attack against the network.