Introduction: Contemporary medicine is undergoing an unprecedented transformation driven by the integration of advanced artificial intelligence (AI) and large language models (LLMs) into clinical workflows. While these technologies objectively enhance diagnostic precision, their implementation creates a fundamental paradox: the increase in technological efficacy often correlates with a decline in patient trust, known as the "AI trust gap." This review examines the tension between algorithmic authority and the necessity of the "human touch," analyzing the impact of digital innovations on clinical autonomy and the patient-physician-AI triad. Materials and Methods: This study presents a detailed analysis of 44 peer-reviewed scientific articles published between 2022 and 2026. The review focuses on Clinical Decision Support Systems (CDSS) across key diagnostic areas, including radiology and pathology. The analysis encompasses the psychological mechanisms of AI acceptance, the risks of automation bias, and the potential of Explainable AI (XAI) to restore clinical transparency. Key Findings: Research reveals that the mere disclosure of AI involvement can reduce patient trust (dropping from 0.50 to 0.30–0.34 in experimental settings). A "paradox of knowledge" was identified, where higher patient literacy regarding AI correlates with increased skepticism. Regarding clinical autonomy, a dichotomy exists: junior clinicians are prone to automation bias, while experts face the risk of "deskilling." The review also discusses the "Algorithmic Consultant" role and the necessity of "Triadic Decision-Making," where AI serves as a transparent partner rather than a black-box authority. Conclusions: The integration of AI requires a reconfiguration of medical practice from a technology-first approach to a human-centered design. Preserving clinical autonomy depends on adopting a "trust but verify" model and implementing XAI strategies to mitigate transparency barriers. The success of algorithmic medicine relies on maintaining the physician's judgment as the cornerstone of care, ensuring that AI functions as a supportive co-pilot.
Kokoszka et al. (Wed,) studied this question.