Animals rely on sensory inputs to navigate complex, dynamic environments. Understanding how the brain integrates these inputs to guide behavior remains a central question in neuroscience. Training AI and animals on the same task offers a unique opportunity to uncover both similarities and critical differences in the strategies used by biological and artificial systems. By pinpointing what the brain accomplishes that modern AI lacks, we can drive advancements in machine learning, pushing AI to more robust, generalizable solutions. In this study, we train mice and reinforcement learning agents on the same virtual open-world foraging task, then compare their performance across a number of visual perturbations to answer a simple question: what can a mouse do that modern AI's cannot? The RL agents process visual input through a two-layer convolutional neural network and learn to intercept a virtual target using proximal policy optimization. Remarkably, both mice and the RL agents learned the task within a few hundred trials and exhibited comparable behavioral patterns, such as similar rotation and body speed as well as a spontaneous development of a random directional bias. However, subtle differences hint at deeper distinctions between biological and artificial navigation strategies and underlying computations. To further probe these differences, we introduced visual perturbations and analyzed their impact on both groups. While both systems adapted to these challenges, the brain's flexibility in processing noisy or ambiguous sensory inputs was significantly better than the RL agent in most cases. Our work lays the groundwork for exploring how differences in sensory processing between the brain and AI systems translate into different behavioral strategies in a visually guided foraging task. In the future, we will investigate the neural representations of task-relevant features in mice and virtual agents, helping us understand how these differences can inform the development of more adaptable AI systems.