Network protection

AI is developing so quickly that we are starting to face problems with breaking encryption keys at all levels.

Many systems were developed before the creation of powerful AI and before of 32 GHZ core CPU that very quickly break almost all known encryption algorithms. We need to think now about how to try to protect yourself. I think this is the threat of our century.

In my opinion, the problem lies in outdated encryption methods, such as MD5, SH-1, SH-2 and many other algorithms that are now vulnerable. And when security systems were built based on them, it seemed that it was practically impossible to find a vulnerability in these encryption methods. These encryption methods become useless, thanks to super powerful processors, which are still being updated and only getting better.

For example, to find the same file that uses MD5 encryption, we need 2^128 degrees power of operations and this can be done in two or three hours on a 32 CPU processor. In the future, the time to bypass this encryption will only decrease and we may face enormous attacks on the servers of well-known corporations.

And now the most interesting thing is the search for encryption algorithms by modern AI. You can try it yourself if you have access to even the legacy ChatGPT-3.5. If you send the AI several keys and ask how they were generated, namely, ask the AI to find what these keys have in common, then the second or third time the AI will tell you what formula these keys were generated using. Wow, it will be fast and I would not say that a person will not be able to cope with it, because a person has cracked not such ciphers, but the time that was spent searching for an algorithm to generate keys is not comparable to the human brain and AI.

What door is AI opening now and where does it lead? We can only hope that the use of AI will be somehow limited and its use for illegal activities prohibited. Yes, this will be a good solution and most will not have access to Pandora's Box. But what about the narrow circle of people for whom such an opportunity will still be available?

Everything that I wrote above concerned security and civilian infrastructure, but a state that owns such AI can use it against its potential enemies, and here we can only imagine how AI will be able to intercept control of UAVs, Satellites and any other weapons, and this is much more serious than we could ever imagine.

Thank you for reading to the end.

Irakli Mekvabishvili, Founder of TOPMADEAPP LLC.

18 of February 2024