Generative AI for 3D Film and Animation Modelling: Pathways, Workflows, and Emerging Standards
Liang Cao
Communication University of China, Nanjing, China.
Jun Dong *
Jiangsu Communication and Media School, Nanjing, China.
*Author to whom correspondence should be addressed.
Abstract
Aims: This study aims to investigate how Generative Artificial Intelligence for Content (AIGC) is transforming 3D modelling from a specialised and labour-intensive process into a semantically driven, collaborative, and accessible paradigm. The objective is to systematically outline its technological evolution, key methods and platforms, and propose an integrated “AI-first, DCC-refined” framework for education and production.
Study Design: This work adopts a comprehensive analytical and comparative study design, combining a systematic review of current AIGC technologies with practical workflow demonstrations and pedagogical integration models. Three major technological pathways—Text-to-3D, Image-to-3D, and Sketch/Language-to-Edit—form the structural backbone of the analysis.
Place and Duration of Study: The study was conducted as part of an ongoing interdisciplinary research project in digital content creation and game technology, the project lasted for more than a year, focusing on digital media and computational design through a number of academic and industrial collaboration projects.
Methodology: The research integrates literature synthesis, workflow mapping, and benchmarking using standard quantitative metrics (CD/EMD/F-score, PSNR/SSIM/LPIPS, and normal consistency). It also designs and validates teachable end-to-end AIGC workflows, course structures, prompt engineering guidelines, industrial implementation manuals, and compliance verification protocols to bridge research, education, and production.
Results: Findings indicate that AIGC-based 3D modelling substantially improves iteration speed, multimodal expressivity, accessibility, and scalability compared with traditional manual workflows. Nevertheless, technical and ethical challenges persist, notably in topological control, semantic alignment, editability, and compliance assurance.
Conclusion: The integration of multi-view diffusion models, large reconstruction networks, and hybrid neural field–to–mesh pipelines—aligned with open standards such as OpenUSD and glTF—is identified as essential for achieving industrial-grade applications. The study contributes an integrated framework that provides a scalable and compliant pathway for embedding AIGC into animation, gaming, XR, and digital twin production ecosystems.
Keywords: Generative AI-empowered, 3D modeling, AIGC tools, pathways, development trends