Background: Generative artificial intelligence (GenAI) is rapidly becoming part of medical students' daily learning. While its benefits for information retrieval and revision are well recognized, less is known about how students perceive its influence on their own clinical reasoning and professional identity. This study explored how medical students balance AI-assisted learning with independent thinking during their undergraduate training. Methods: A qualitative descriptive study was undertaken involving undergraduate medical students with regular experience using GenAI for academic purposes. Participants engaged in moderated focus group discussions examining AI use, decision-making, learning habits, confidence, and concerns regarding future clinical practice. Audio recordings were transcribed and analyzed using conventional content analysis to identify recurring patterns and educational implications. Results: Students described AI as a valuable learning companion but expressed concern about becoming "too comfortable" relying on generated responses. Four categories emerged: AI as a cognitive shortcut, confidence without competence, protecting independent clinical reasoning, and defining the physician's unique role. Participants emphasized that while AI accelerated learning, excessive dependence reduced opportunities to struggle with complex problems, justify clinical decisions, and develop diagnostic reasoning. Many advocated for explicit guidance on responsible AI use rather than restrictions on access. Conclusion: Medical students do not view AI as a replacement for clinical thinking but seek educational frameworks that help them use it responsibly. Integrating AI literacy with structured reasoning activities may enable learners to benefit from technological advances while preserving the analytical skills and professional identity expected of future physicians.