A landmark event has occurred in the field of cybersecurity: Anthropic's AI, Claude, has discovered previously unknown vulnerabilities in two encryption systems. These are mathematical algorithms that protect our messages, logins, and banking data from outsiders.

It is important to emphasize: the current encryption standards used in everyday life have not been compromised. However, the significance of this discovery extends far beyond a simple bug fix. The machine found what leading global cryptography experts had overlooked.

Vulnerability in HAWK: Key at Risk

HAWK is a digital signature, a kind of identity verification for websites and messages. The system was originally designed with the threat of quantum computers in mind. For two years, experts found no flaws in it, but Claude identified a hidden pattern in HAWK's mathematical foundation in just 60 hours.

According to calculations, the smallest HAWK key can now be cracked in 2^38 steps instead of 2^64. This means a reduction in required computing resources by approximately 67 million times. There is a solution to the problem, but it is unpleasant: the key size must be doubled, which completely negates HAWK's main advantage.

Test on AES: "Möbius Bridge"

AES (Advanced Encryption Standard) is the primary standard that has encrypted most internet traffic since 2001. Anthropic did not manage to crack full AES. Claude analyzed a simplified version of the algorithm—with 7 rounds instead of the standard 10.

The AI found a non-standard attack method, which it named "Möbius Bridge." The trick's essence is that one stage of the selection process turned out to be redundant. As a result, the most effective attack was accelerated by 200–800 times. Notably, Claude worked alone for three days, and then two specialists spent nearly a month double-checking its response.

What is the significance of this work?

The main delay now occurs at the stage of verifying the result. Anthropic notes that humans can no longer keep up with the model's pace. This is not the first time errors have been found only at late stages. In 2022, researchers managed to crack another post-quantum algorithm, SIKE, in about an hour on an ordinary laptop.

It is important to understand: Bitcoin (BTC) and Ethereum (ETH) operate on a different principle, and these attacks do not affect them. However, time does not stand still. Google plans to transition to new standards by 2029, and U.S. government agencies have been given a quantum deadline of 2031.

Now, NIST 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. In the near future, the company plans to hold a special seminar with scientists.

My expert commentary: This event marks a new stage in cryptography. AI is not just automating routine tasks but is beginning to act as a full-fledged researcher capable of finding non-trivial solutions. The question "should the security check of your data be left to artificial intelligence?" is no longer hypothetical—it is our reality, and the industry will have to adapt to it.