Quantum-Resistant Crypto Algorithm HAWK Found Broken After Years of Use

Text
The U.S. National Institute of Standards and Technology (NIST) has removed the HAWK quantum-resistant cryptography algorithm from its post-quantum cryptography (PQC) standardization process after a flaw was identified by Anthropic’s security model, Mythos. The algorithm, designed to protect digital communications from future quantum-computer attacks, had progressed through two rounds of NIST’s evaluation before entering the third round, which was intended to detect vulnerabilities like the one uncovered. Subheading
HAWK’s Role in Quantum-Resistant Security
HAWK, a digital signature scheme, was among the candidates under consideration for a U.S. federal standard to safeguard encrypted data against potential quantum-computer threats. Quantum computers, once realized at scale, could break widely used cryptographic systems like RSA and ECC by efficiently solving mathematical problems these systems rely on. The third round of testing, which HAWK was undergoing, aimed to simulate real-world attack scenarios and identify weaknesses that might have been overlooked in earlier stages. Anthropic’s Mythos model, a security tool designed to analyze cryptographic protocols, detected a critical vulnerability in HAWK’s design. While the exact nature of the flaw remains unspecified in public reports, the developer confirmed it rendered the algorithm insecure, prompting its withdrawal. The third round, in particular, is designed to identify subtle vulnerabilities that might not surface in earlier stages. HAWK’s withdrawal underscores the challenges of developing cryptographic systems that can withstand both current and future threats. While practical large-scale quantum computers remain years away, governments and private entities are proactively preparing for potential threats. HAWK’s removal does not derail the PQC initiative but serves as a reminder of the complexities involved in securing digital infrastructure. Subheading
What Comes Next? NIST is expected to continue its third-round evaluations, with results likely to shape the final selection of PQC standards. Anthropic’s role in identifying the flaw also highlights the growing intersection of AI and cybersecurity, where machine learning tools are increasingly used to detect vulnerabilities in complex systems.