In a pre-post experiment (n = 41), we test the impact of an AI Coach's explanatory communications modeled after the instructions of human driving experts. Participants were divided into four (4) groups to assess two (2) dimensions of the AI coach's explanations: information type ('what' and 'why'-type explanations) and presentation modality (auditory and visual). We directly compare how AI Coaching sessions employing these techniques impact driving performance, cognitive load, confidence, expertise, and trust in an observation learning context. Through interviews, we delineate the learning process of our participants. Results show that an AI driving coach can be useful for teaching performance driving skills to novices. Comparing between groups, we find the type and modality of information influences performance outcomes. We attribute differences to how information directed attention, mitigated uncertainty, and influenced overload experienced by participants. These, in turn, affected how successfully participants were able to learn. Results suggest efficient, modality-appropriate explanations should be opted for when designing effective HMI communications that can instruct without overwhelming. Further, they support the need to align communications with human learning and cognitive processes. Results are synthesized into eight design implications for future autonomous vehicle HMI and AI coach design.