Artificial intelligence (AI) and machine learning describe an approach where software can be trained to perform tasks and recognizing patterns, such as against a set of images. Wildlife camera traps are placed in the field near and far from infrastructure, such roadway bridges, and are essential equipment to monitor animal movement and occupancy. Large arrays (dozens to hundreds of cameras) result in large number of images (hundreds of thousands), setting up a ripe opportunity to apply AI techniques to help identify animals within the digital photographs. Rapidly and accurately processing images through most workflows can involve a lot of staff time and potentially result in data transcription and other errors. For this investigative method to be efficient and accurate at large scales, it is necessary to automate certain steps to reduce labor and increase accuracy and consistency. In this study, we evaluated Microsoft MegaDetector which detects animal presence in images. We compared version 3 with version 5 of MegaDetector due to changes in the underlying AI technology. MegaDetector v3 uses Faster-RCNN object-detection system with an InceptionResNetv2 base network, whereas MegaDetector v5 uses You Only Look Once v5, which is based on DarkNet and PyTorch frameworks and built in data augmentation processes to improve trainability. We describe the AI Image Toolkit (AIT, https://ait.dudek.com), a web-based system incorporating MegaDetector v5 and designed to identify those camera trap images containing animals (>95% accuracy). A single instance of AIT is capable of accurately classifying animal-containing images at a rate of ~50,000 images/week. Ten cloned instances on a single device would increase this rate to 500,000 images/week, sufficient to process “full” SD cards from about 25 cameras/week. The AI sub-system transfers animal-containing images to the camera trap image management system where they are organized and tagged with additional information, including species, gender, age class, and more. These data are compiled within the system and provide data summaries by project or by camera location. The data and metadata can be queried and automatically packaged into formats used in GIS or statistical analysis; for example, occupancy models, diversity indices, and effectiveness of crossing structures.