This paper examines some recent developments in artificial intelligence (AI)-based cryptography. In particular, it takes into account the use of evolutionary computing (EC) and machine learning (ML) for data encryption and analysis. An introduction to Artificial Neural Networks (ANNs) and the fundamentals of Deep Learning with Deep ANNs is provided. The study examines two topics in this regard: (i) the use of EC and ANNs to create distinct and unclonable ciphers; and (ii) machine learning techniques to determine whether or not finite binary strings are truly random for use in cryptanalysis.