Abstract Black carbon (BC)‐containing particles are the most strongly absorbing component of atmospheric aerosol, contributing to global warming. Quantifying their radiative effects remains challenging because BC absorption depends on their mixing state which can evolve through atmospheric aging. This study presents unique direct aircraft observations of BC absorption enhancement and mixing state in urban outflows from Paris. We show that the absorption of BC either increases or remains unchanged during plume transport at the regional scale. By using an unsupervised machine learning approach, we identify clusters of BC properties associated with specific meteorological conditions and chemical composition. The analysis reveals that relative humidity (RH) significantly influences BC absorption enhancement. We derive a time‐dependent rate of 0.20 hr−1 for BC absorption enhancement at RH above 60%, decreasing the rate to less than half in drier conditions. These findings emphasize the need to include RH‐dependent parametrizations to improve estimates of BC radiative effects.