Artificial intelligence adoption in the cultural heritage sector: A systematic review of trends, application domains, and challenges

The adoption of Artificial Intelligence (AI) in the cultural heritage sector has increased significantly in recent years, driven by its potential to support heritage preservation, institutional management, and visitor engagement. Despite this growing interest, existing research remains fragmented and lacks a comprehensive synthesis of research trends, application domains, and adoption challenges. To address this gap, this study conducts a systematic literature review of 64 peer-reviewed articles published between 2020 and 2024 following the Kitchenham guidelines. The review analyzes the current landscape of AI adoption in cultural heritage by examining research trends, major application domains, and recurring challenges reported in the literature. The findings identify five major application domains of AI in the cultural heritage sector: restoration and preservation; documentation and digitization; management and decision support; generation and creative reconstruction; and education and experience. Among these domains, education and experience emerge as the most extensively explored areas, reflecting the growing emphasis on AI-enabled personalization and interactive visitor engagement. The analysis further reveals that current research predominantly focuses on technical development and proof-of-concept systems, while empirical evaluation within operational cultural heritage institutions remains limited. Notably, the findings underscore that human behavioral factors — including user acceptance, professional and public trust, and institutional readiness — play a critical role in shaping AI adoption outcomes, highlighting the importance of human-centered perspectives in the design and evaluation of AI systems for cultural heritage. In addition, the review identifies five major categories of challenges affecting AI adoption: data availability and quality; technical and operational; ethical, privacy, and ownership; bias, trust, and validation; and collaboration and interdisciplinary. This study employs the Technology–Organization–Environment (TOE) framework as the central analytical lens to systematically interpret and structure AI adoption challenges in the cultural heritage sector — a domain that has received limited attention in TOE-based adoption research. The TOE-based analysis reveals that AI adoption in this sector is shaped not only by technological factors but also by organizational readiness and environmental contexts. To further examine the distribution of challenges across domains, this study develops a challenge–application matrix that reveals domain-specific challenge profiles, and synthesizes the findings into a TOE-based thematic framework that maps challenge intensity across the five application domains. As its central contribution, this study positions the TOE framework as an integrative lens for synthesizing AI adoption challenges in the cultural heritage sector, providing a structured analytical foundation that bridges fragmented findings and highlights critical gaps in the literature. The findings suggest that future research should prioritize empirical studies on organizational adoption, governance frameworks, and user perspectives to support the contextually sensitive and institutionally grounded adoption of AI technologies in the cultural heritage sector.