SIKE Protocol Vulnerable to AI-Driven Attacks

• •The SIKE protocol was broken in one hour using conventional computing.
• •AI's ability to connect obscure mathematical ideas poses new risks to cryptography.
• •Organizations must prepare for the potential failure of newly implemented cryptographic algorithms.

The Supersonic Multivariates Isogeny Key Encapsulation (SIKE) protocol, a candidate for quantum-resistant encryption, was compromised in one hour using a standard laptop. Mathematicians Wouter Castryck and Thomas Decru linked SIKE's structure to a theorem by Ernst Kani from 1997, allowing them to recover a private key for the smallest parameter set. This incident raises concerns about the ability of frontier AI systems to rapidly identify and exploit cryptographic weaknesses. The attack did not require quantum computing, demonstrating vulnerabilities in conventional systems. As AI continues to advance, the window for discovering and exploiting weaknesses in cryptographic algorithms is shrinking. Organizations must not only replace vulnerable algorithms but also prepare for potential failures of replacements. The recent successes of AI in solving long-standing mathematical problems further underline the urgency for reevaluating cybersecurity strategies.