Claude, an artificial intelligence developed by Anthropic, has achieved a breakthrough in the field of cryptanalysis. During research, the model discovered previously unknown vulnerabilities in two encryption systems that protect messages, logins, and banking data from outsiders. These are the HAWK digital signature protocol and a simplified version of the AES standard.

It is important to emphasize: the current algorithms used in everyday life remain secure. However, the news itself is significant for another reason — the machine noticed what leading global experts had missed over years of work.

Vulnerability in HAWK: key weakened by 67 million times

HAWK is a cryptographic protocol designed to protect against attacks from quantum computers. For two years, experts found no flaws in it. Claude needed only 60 hours to identify a hidden mathematical pattern at its core.

According to calculations, cracking the smallest HAWK key is now possible in 2^38 steps instead of 2^64. This means computational costs have been reduced by approximately 67 million times. A solution exists, but it is unpleasant: the key size must be doubled. As a result, HAWK's main advantage — compactness — is lost.

Attack on AES: "Möbius Bridge" accelerates cracking by 800 times

AES (Advanced Encryption Standard) is the primary encryption standard protecting most internet traffic since 2001. Full AES remained untouched, but Claude analyzed a simpler version of the algorithm with 7 rounds instead of the standard 10.

The AI found a non-standard attack method called the "Möbius Bridge." The trick's essence is that one stage of selection turned out to be redundant. As a result, the most effective attack was accelerated by 200–800 times. It took Claude three days to complete this work alone. Then two specialists spent nearly a month double-checking the obtained answer.

Expert Opinion

The main delay now occurs at the stage of verifying results. Anthropic notes that humans can no longer keep up with the model's pace. This is reminiscent of the story of the post-quantum algorithm SIKE, which was cracked in an hour on a regular laptop in 2022 — at the late stages of standardization. Now NIST must evaluate a public attack on an algorithm that was still being considered for implementation. Anthropic indicated that not all attack results have been published yet, and the company plans to hold a special seminar with scientists in the near future.

My analysis: This case is a clear signal for the entire industry. AI is beginning to play a role not just as a tool, but as an active participant in security auditing. Cryptographers will have to adapt to a new reality where the time to find vulnerabilities shrinks to days, not years. Should the security verification of your data be entrusted to artificial intelligence? The question is not whether it is possible, but whether we are ready for it.