Influenza A virus (IAV) remains a major cause of respiratory infections, resulting in significant global morbidity and mortality. Understanding the host immune response is key to elucidating disease mechanisms, requiring human-relevant model systems that preserve the complex lung microenvironment. Precision-cut lung slices (PCLS) represent such a model, maintaining native architecture and cell–cell interactions, thereby enabling ex vivo infection studies under near-physiological conditions. However, current optical imaging approaches for PCLS─such as conventional two-photon excited fluorescence (TPEF) microscopy─primarily reveal structural information but lack the metabolic and functional contrast needed to differentiate immune cells, such as macrophages, from surrounding lung tissue. To overcome this limitation, we applied two-photon fluorescence lifetime imaging microscopy (2P-FLIM) to IAV-infected human PCLS, capturing intrinsic autofluorescence signals (mainly NAD(P)H) that report on cellular metabolism. This label-free approach enables discrimination of macrophages within tissue, but time-intensive acquisition and analysis limit throughput for systematic studies. To accelerate image collection and interpretation, we trained a neural network on annotated 2P-FLIM data sets, where fluorescence intensity images were labeled via phasor-based lifetime segmentation. The model learned morphological and intensity cues to predict the macrophage distribution directly from intensity-only z-stacks, enabling rapid segmentation. This AI-enhanced FLIM framework bridges physiological relevance and analytical efficiency─combining human PCLS, phasor-based 2P-FLIM, and deep learning to achieve fast, label-free visualization and quantification of pulmonary macrophages in a human model in one-fifth of the time compared to only 2P-FLIM. The approach highlights the potential of integrating advanced photonics and AI for high-throughput immunometabolic studies and antiviral research.