The Rise of Intelligent Cinema: Systematic Mapping of Natural Language Processing Techniques in Film Analysis, Translation, and Recommendation

This study presents a PRISMA-guided systematic literature review of 171 peer-reviewed journal articles published between 2021 and 2025, examining the application of Natural Language Processing (NLP) in the film industry. The objective is to map methodological trends, identify research gaps, and propose a unified analytical framework across five major domains: sentiment and emotion analysis, content generation, multilingual translation, recommendation systems, and knowledge extraction. The research contribution is the development of an integrated “cinematic intelligence” framework that synthesizes previously fragmented NLP applications within a single domain-specific mapping study. Studies were selected from the Scopus database based on predefined inclusion criteria (English-language, peer-reviewed journals, film-related NLP focus), ensuring methodological transparency and reproducibility. The findings indicate a clear transition toward transformer-based, multimodal, and explainable AI architectures. Transformer models such as BERT, GPT, and LLaMA dominate emotion analysis, automated script generation, and cross-lingual subtitle processing. However, persistent challenges include limited long-context narrative modeling, cross-cultural generalization constraints, cold-start bias in recommendation systems, and the absence of standardized evaluation metrics for AI-generated creative content. Overall, the review demonstrates that NLP in cinema is evolving toward integrated, scalable, and context-aware systems, while highlighting critical research opportunities and practical implications for both academic and industry stakeholders.