The world of cryptography has faced an unexpected challenge: artificial intelligence, specifically the Claude model from Anthropic, has discovered critical vulnerabilities in two encryption systems. This is not about breaking existing algorithms that protect bank transactions, logins, and messages, but about identifying mathematical weaknesses that had previously eluded leading experts. This event overturns the traditional notion of who and how should verify the reliability of cryptographic standards.
HAWK Protocol: 67 Million Times Faster to Break
One of the "victims" of the AI analysis was the HAWK protocol—a digital signature designed with the threat of quantum computers in mind. For two years, specialists meticulously studied its mathematical foundation and found no flaws. However, Claude needed only 60 hours to uncover a hidden pattern. According to Anthropic, breaking the smallest HAWK key is now possible in 2^38 steps instead of the expected 2^64. This means the computational cost of an attack has been reduced by approximately 67 million times. Unfortunately, the solution is unpleasant: the key size must be doubled, which negates the protocol's main advantage—its efficiency.
AES: "Möbius Bridge" Accelerates Attack by Hundreds of Times
The second object of study was AES (Advanced Encryption Standard), the gold standard of encryption that has protected most internet traffic since 2001. Anthropic did not manage to break the full 10-round AES. Instead, Claude analyzed a simplified version—with 7 rounds. Here, the AI found a non-standard approach, dubbed the "Möbius Bridge." The essence of the trick is that one of the key selection stages turned out to be redundant. As a result, the most effective attack was accelerated by 200 to 800 times.
Claude spent three days working autonomously on this task. After that, two specialists spent nearly a month double-checking the results. This vividly demonstrates a new reality: humans can no longer keep up with the pace of AI work.
Why This Matters for Cryptocurrencies and Beyond
The main delay now occurs at the stage of verifying results. Anthropic directly states that such cases are not isolated. In 2022, researchers managed to break another post-quantum algorithm, SIKE, on an ordinary laptop in about an hour. This underscores the fragility of many modern cryptographic solutions.
It is important to note: Bitcoin (BTC) and Ethereum (ETH) operate on different principles, and these attacks do not directly affect them. However, time is running out. Google plans to transition to new standards by 2029, and U.S. government agencies have a quantum deadline of 2031. NIST now must evaluate a public attack on an algorithm that was still being considered for implementation. Anthropic has indicated that not all attack results have been published yet, and the company will soon hold a special seminar with scientists.
My expert opinion: This breakthrough is not a reason for panic, but a powerful signal for the industry. AI is becoming not just a tool, but an independent researcher capable of finding what humans miss. The question now is not "can AI break encryption?" but "how quickly can we adapt our security system to this new reality?" Should the security of your data be entrusted to artificial intelligence? It seems we have no other choice.