Answers are grounded in this paper and cite their sources.
The rapid integration of artificial intelligence (AI) into surgical practice is transforming clinical decision-making processes, raising critical questions regarding trust, autonomy, and responsibility in the operating room. This systematic literature review aims to synthesize current research on AI-driven surgical decision-making, with a particular focus on human–AI interaction and its ethical, cognitive, and legal implications. The review was conducted in accordance with PRISMA guidelines. A comprehensive search of Scopus, Web of Science, PubMed, and Google Scholar identified a total of 312 records published between 2015 and 2025. After the removal of duplicates and screening based on predefined inclusion and exclusion criteria, 47 studies were included in the final analysis. These studies were categorized into three main thematic domains: trust in AI systems, autonomy and human oversight, and responsibility and accountability. The findings indicate that AI technologies significantly enhance predictive accuracy, intraoperative support, and clinical decision-making efficiency (Gao et al., 2021; Shao et al., 2023). However, their increasing integration introduces challenges related to automation bias, reduced clinical autonomy, and ambiguity in legal responsibility ((O’Sullivan et al., 2019; Morris et al., 2023)). Trust in AI emerges as a multidimensional construct influenced by transparency, explainability, and user experience (Lopes et al., 2025). The review highlights the need for improved explainability, standardized validation frameworks, and clearly defined accountability structures. It also identifies key gaps in the literature, particularly the lack of empirical studies on real-time human–AI decision-making and long-term impacts on surgical training. Future research should adopt interdisciplinary approaches to ensure the safe, ethical, and effective integration of AI into surgical practice.
No takes yet. Share an insight, caveat, or question.
Marzec et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: