Abstract: Grand challenges in nuclear physics require a paradigm shift in nuclear many-body theory that can be accomplished by introducing artificial intelligence (AI) techniques to accurately solve the structure of large nuclei and their interactions with neutrinos.We introduced an AI representation of the imaginarytime propagator that enables quantum Monte Carlo studies of medium-mass nuclei being experimentally investigated at both the Argonne Tandem Linac Accelerator System and the Facility for Rare Isotope Beams.Concurrently, we developed novel deep-learning methods to accurately reconstruct the nuclear electroweak response functions of atomic nuclei, a critical ingredient for the success of the Deep Underground Neutrino Experiment, DOE's flagship neutrino-oscillation experiment in which Argonne aims to play a growing role.