Machine Learning Based Encrypted Content Type Identification

In the advancing era, Machine Learning has become the backbone of IT and is being used almost in every system. Whereas Cryptography is another widely used technology, which is used for the communication of data via secure means. If the cryptosystems provide indistinguishability, it is considered secure, which means that the attacker cannot get anything from encrypted data, in case of chosen ciphertext attack. To check the feasibility of distinguishability on the ciphertext of secured block ciphers and the identification of the underlying content, this research has applied cryptanalysis on AES-128, CBC, and ECB mode over multiple ML classification models of SVM, KNN, and RF. Datasets were created by using the frequency distribution method, and they were divided into training and testing datasets. The results demonstrate that the EBC mode of AES 128 encryption for different data types is found susceptible to content identification while the accuracy for data encrypted by AES 128 in CBC mode remains low to yield any information regarding its content type.