This research addresses the challenge of privacy in 6G wireless networks using Artificial Intelligence. Privacy and prevention of eavesdropping are significant concerns in 6G networks, warranting motivation for new approaches, such as Physical Layer Security. Physical Layer Security, applied at Layer 1 of the OSI stack, provides low-latency privacy for over-the-air waveforms at the transmit antenna. This research describes a new application of Key Based Physical Layer Security (KB-PLS), capable of mapping secret information bits onto MIMO precoders derived from codebooks. This work defines codebook precoder classification as a problem that can be solved using neural networks. We describe a communication protocol in which Singular Value Decomposition (SVD) recovers precoder values. The receiver applies machine learning to recover bits by matching noisy precoders to specific codebook entries. Aided by accelerating technologies such as NVIDIA GPUs and cuDNN libraries, machine learning methods are discussed that replace and improve upon existing mathematical derivations of the 6G Privacy protocol.