New Approaches to Wildlife Detection tackles the fundamental yet intractable problem of how might we more cost-effectively obtain accurate, precise, and simultaneous data on multiple wildlife populations at the scale of the Boreal Plains and Shield ecozones to monitor complex population dynamics and test ecological theory. While data on relative species abundances at a local scale can be obtained from both traditional ecological knowledge (TEK) and science (e.g., surveys, mark-recapture analysis), when scaled-up to the extent of an ecozone, data equivalencies fall apart. Our partnership is positioned to provide the food-web dynamics modelling that has been beyond the reach of ecologists due to lack of data on densities of interacting species, especially for species that are costly to monitor like large mammals in forested environments. By working with a multi-spectral, high resolution but scalable imaging payload (up to 3-cm colour plus co-boresighted infrared; coverage up to 1000 km2 per day using fixed-wing aircraft) we can remotely census large mammals using artificial intelligence (AI) and deep learning to optimize manual identification and counting in complex environments.