Purpose Generative AI has challenged traditional written assignments while creating opportunities for interactive, formative learning. In Human-Microbe Interactions, second-year medical students traditionally completed case-based essays. We piloted AI chat simulations as an alternative to capture learner reasoning, provide individualized feedback, and support clinical application. Methods During Fall 2025, students completed weekly essays or AI chat simulations. Simulations were built in Hofstra University's closed ChatGPT Edu environment using existing prompts and answer keys. In 20-minute cases, students interacted with a simulated attending physician, generated differentials, requested tests, interpreted findings, and responded to prompts aligned with basic science objectives. The AI provided feedback and summarized strengths and knowledge gaps. Students submitted transcripts through Canvas for pass/fail credit. Evaluation included participation data, Likert-scale items, and qualitative feedback. Results Ninety-seven students completed the survey (96% response rate). Fifty used the AI simulation at least once; 40 selected the simulation instead of the essay at least three times. Simulation users reported that the activity helped them identify gaps in understanding (M=4.4/5), apply knowledge rather than recall facts (4.3/5), remain engaged (4.3/5), demonstrate understanding as effectively as essays (4.3/5), and want to use simulations in future courses (4.3/5). Qualitative feedback emphasized immediate feedback, reasoning practice, and active learning. Conclusions AI chat simulations may offer a feasible, scalable alternative to written assignments by supporting formative feedback, engagement, and clinical reasoning practice. This pilot suggests that scaffolded conversational AI can help redefine competence by emphasizing application, adaptability, and reflective learning while preserving faculty oversight and course alignment.