Artificial intelligence in NM-protein interactions. The future translational applications of nontechnology in human body will greatly be influenced by NM-protein interactions. Exploring the instinctive and automatic prediction of nanosurface interactions with protein using machine learning will help in estimating the risk profiles of using NM and also the feedback can guide us designing the desired NM that facilitates beneficial interactions. The unique NM features ((a) size, shape, granularity, pH, etc.) that are hypothesized to contribute to NM-protein interactions could be considered as (a nonrigid set of) “input feature vectors” restoring the knowledge from available experimental data (b). The set of features can then be trained against appropriate target functions to build machines (c). Different combination of features needs to be attempted till optimization. This could give us a hypothesis as to whether or not (or in what extent) the NM will interact with specific/nonspecific proteins (d). A successful endeavour of this exercise should further be taken into the prediction and design of beneficial nanoprotein interactions. It should be noted with care that the training and cross-validation of such machine learning based predictors is a long-term exercise falling under continuous evaluation (on new and updated datasets as they get available) demanding an ever-increasing accuracy until it is “adequate” (e).