Automated Detection of Encryption Algorithms Using AI Techniques

In today's digital era, cryptographic algorithms play a vital role in safeguarding sensitive information, maintaining confidentiality, and defending against cyberattacks. With the growing variety of encryption methods and modes, identifying the specific algorithm used to generate a ciphertext has become increasingly challenging. To address this, we propose a machine learning solution that automatically identifies cryptographic algorithms using the Random Forest classifier. The system processes a vast dataset called “encrypted-data,” which contains 39 different algorithms and modes, by extracting essential features like byte frequency and entropy from the ciphertexts. These features form a 257-dimensional Feature vector used for classification. Random Forest models are trained with these data. During testing, models are loaded is used to predict the encryption algorithm for a given ciphertext. Our method offers a scalable and efficient approach to cryptographic identification, surpassing traditional techniques.