Artistic Creation Practice and Exploration Based on AI Human-Machine Co-creation and Virtual Simulation

Research Paper This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. DOI: 10.5281/zenodo.17852678 Page 14 1. Introduction 1.1 Research Background and Motivation We are living in an era profoundly reshaped by digital technologies. In recent years, cutting-edge technologies such as generative artificial intelligence (AIGC), virtual simulation, and spatial computing have experienced explosive growth. This not only presents unprecedented opportunities and challenges across industries but also injects powerful transformative momentum into education, culture, and creative industries [1]. At the macro level, national strategies emphasize "vigorously advancing the modernization of industrial systems and accelerating the development of new-quality productive forces," with the core driving force being emerging growth models centered on the digital economy and future industries [2]. Against this backdrop, fields like digital media art, film and television art, and game art—key engines of digital economic development—are undergoing rapid evolution in their creative paradigms, production tools, and talent requirements. In the realm of artistic creation, while traditional approaches have established a solid artistic foundation, they increasingly reveal limitations when addressing complex project demands, pursuing peak efficiency, and pushing creative boundaries. Artists and students frequently encounter three major challenges: First, the gap between creative conception and visual realization—transforming fleeting inspirations into detailed concept sketches often requires significant time and effort. Second, the technical barriers and implementation costs are substantial, particularly for 3D scenes, grand narratives, or specialized visual effects, which demand advanced software skills and robust hardware support. Third, the Artistic Creation Practice and Exploration Based on AI Human-Machine Cocreation and Virtual Simulation Li Xinyi GRADUATE UNIVERSITY OF MONGOLIA, Ulaanbaatar, Mongolia. Corresponding Author: Li Xinyi This study investigates innovative art creation methodologies leveraging AI-driven human-machine collaboration and virtual simulation technologies, with accessory design serving as a practical case study. The research establishes a closedloop workflow model encompassing "Concept Generation-Asset Transformation-Immersive Tuning-Iterative Optimization" (CAII). Through the implementation of the "Time-Space Ring" buckle design—from AI-generated concepts to 3D modeling and dynamic validation with leather bags in virtual environments—the methodology successfully bridges the gap between static renderings and interactive prototypes, significantly enhancing design efficiency and feasibility. Practical applications demonstrate that this approach not only overcomes traditional design constraints in functional verification and user experience evaluation, but also provides a forward-looking digital solution for addressing the rapidly evolving demands of the fashion industry KEY WORDS: AI collaboration, virtual simulation, accessory design, AIGC, workflow Abbreviated Key Title: UAI J Arts Humanit Soc Sci ISSN: 3048-7692 (Online) Journal Homepage: https://uaipublisher.com/uaijahss/ Volume2 Issue12 (December) 2025 Frequency: Monthly Research Paper This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. DOI: 10.5281/zenodo.17852678 Page 15 trial-and-error costs in the creative process are high, as adjustments to materials, lighting, and composition typically involve repeated rendering and lengthy iterations, thereby restricting artistic exploration. Meanwhile, the maturation of AI-driven human-machine collaboration and virtual simulation technologies has opened new avenues to address these challenges. AIGC technology can transform textual descriptions into high-quality visual images at unprecedented speed, dramatically expanding creative scope and visualization efficiency [3]. Virtual simulation technology, by creating highly realistic and interactive 3D environments, enables creators to "immerse" themselves in their works for immersive construction, review, and modification—marking a paradigm shift from "2D plane drawing" to "3D spatial modeling." Therefore, exploring how to organically integrate these two technologies to establish an efficient, intuitive, and creative new methodology for artistic creation not only addresses urgent practical needs but also holds profound developmental potential. This represents a crucial endeavor to respond to contemporary demands and empower creative productivity. 1.2 Review of Domestic and Foreign Research In international research, the integration of AI and art has evolved from early algorithmic art to current human-machine collaboration based on deep learning. Platforms like OpenAI's DALL-E series and Midjourney have been widely adopted in commercial art and conceptual design, with research focusing on prompt engineering, style transfer, and AI's role in creative workflows [4]. Meanwhile, virtual simulation technology has become standard practice in film pre-production and architectural visualization. Disney's StageCraft virtual production system, for instance, achieves real-time scene rendering through massive LED screens, blurring the boundaries between pre-production and post-production [5]. However, existing studies often treat AI-generated content and virtual simulation as separate tools, with insufficient exploration into how they can form a deeply integrated closed-loop workflow in individual or small-tomedium team creations. Domestic research trends: With the government's policy support for the digital economy and metaverse industries, related studies have flourished. The academic community has extensively discussed the artistic value of AIGC, copyright issues, and educational applications. In virtual simulation fields, domestic universities and research institutions have achieved remarkable results in virtual experimental teaching and digital museum construction. However, existing research has two notable shortcomings: First, most studies focus on technical descriptions or single-technology applications, lacking systematic analysis and practical verification of the collaborative mechanisms between AI and virtual simulation in the entire art creation process from a "methodology of creation" perspective. Second, research perspectives tend to emphasize macrolevel educational or industrial applications, while specific pathways for individual artists or student groups to enhance core creative abilities through these technologies remain relatively underexplored. Therefore, this study aims to address the limitations of existing research. Rather than merely introducing techniques, it focuses on exploring the integration pathways and effectiveness of AI-human co-creation and virtual simulation production in artistic practice. By constructing concrete workflow models and validating them through practical cases, the study seeks to provide a new, actionable paradigm for digital art creation. Against this backdrop, this study systematically explores and establishes an AI-driven collaborative art creation methodology that integrates human-machine co-creation with virtual simulation technologies, with feasibility validated through case studies. The significance lies in three dimensions: Theoretically, it provides fresh perspectives for digital art creation theory by examining inspiration generation and aesthetic decision-making in human-machine collaboration. Practically, it develops replicable workflows to reduce technical barriers, enhance creative efficiency, and enable artists to focus on artistic expression. For the cultural and creative industries, it offers actionable insights for adapting to new productive forces, achieving cost-efficiency improvements, and fostering innovative forms. 2. Core Method: AI-Virtual Simulation Fusion Workflow To systematically integrate AI's generative capabilities with virtual simulation validation, this study developed a closed-loop workflow model named "Concept-Asset-Immersion-Iteration" (CAII). The model abstracts the creative process into four core stages, emphasizing rapid cycles and feedback mechanisms to enhance the efficiency and quality of the transformation from concept to final product. (1) Concept Generation Layer: In this phase, generative AI acts as an 'inspiration accelerator.' Researchers utilize a series of prompts to explore styles and conduct visual brainstorming, rapidly generating numerous concept sketches to establish the overall aesthetic direction and atmosphere of the creative work. (2) Asset Conversion Layer: This layer's core task is to convert 2D concept art into 3D assets usable by the virtual engine. Through AI texture generation and model optimization, it transforms selected visual styles into digital models, materials, and textures, building the foundational components of the virtual world. (3) Immersive Editing Layer: After assets are imported into the real-time rendering engine, creators can enter the virtual scene via VR devices for immersive review and editing. This layer reveals spatial, proportional, and lighting issues that are hard to detect on 2D screens, and supports real-time "what you see is what you get" modifications. (4) Iterative Optimization Layer: This is the key to achieving a "closed loop" in the model. Issues identified during immersion tuning (such as material inconsistencies) are fed back to the AI as new inputs for targeted regeneration or optimization, initiating a new round of fine-tuning until the creative goal is achieved. The core strength of this CAII model lies in its dynamic iterative nature, seamlessly integrating AI's' broad exploration 'with virtual simulation's 'deep validation' to create a spiral upward creative process, offering clear methodological guidance for practical applications. Chapter 3 Case Study: Virtual Accessory Design and Matching Validation Based on CAII Workflow To validate the practicality and effectiveness of the CAII workflow model proposed in Chapter 2, this study selects a virtual accessory with retro-futuristic aesthetics—the "Time Ring" buckle—as a creative case. This case aims to comprehensively demonstrate the entire process from Research Paper This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. DOI: 10.5281/zenodo.17852678 Page 16 concept generation, 3D modeling, material rendering to dynamic matching validation with virtual leather bags, fully showcasing the closed-loop advantages of virtual simulation technology in accessory design. 2.1 Theme Design and AI Concept Generation This case study focuses on "retro-futuristic fasteners," blending Victorian ornate aesthetics with sci-fi mechanical designs. In the "Concept Generation Layer" of the CAII workflow, we utilize Jimo for rapid visual exploration. Prompt word engineering is key. The initial broad prompt word is:a intricate belt buckle, steampunk style, bronze and copper, glowing cyan elements, hyper-detailed, 4k --v 6.0.The generated result has a steampunk style but is overly mechanical and lacks retro charm. After several iterations, the optimized prompt is:a Victorian-era ornamental buckle, with delicate floral engraving, but with a futuristic core of glowing light and precise mechanical joints, made of patina bronze and polished brass, photorealistic, studio lighting.The results of this round successfully balanced retro and futuristic elements, and were selected as the 'concept anchor point' to provide precise visual references for subsequent 3D modeling. Figure 1 Dreamina AI Image Note: A Victorian-era ornamental buckle is sourced from the Dreamina AI 2.2 Full Process of Accessories Production under Virtual Simulation This phase corresponds to the 'Asset Transformation Layer' and 'Immersive Tuning Layer' in the CAII workflow, replicating the entire process from modeling to final rendering of accessories in 3D software (e.g., Blender) and real-time rendering engines. 1. High-Precision 3D Modeling and Structural Simulation: Based on the AI concept design, we conducted highprecision 3D modeling in Blender. The focus was not only on visual fidelity but also on simulating functional mobility. The "Ring of Time and Space" buckle was designed with a mechanical opening mechanism. Using Blender, we implemented drive joints for the buckle's opening action. By configuring a simple rotation controller, we could simulate the complete opening animation in virtual environments, allowing us to verify the structural rationality and usability during the design phase. 2. Programmatic Materials and Physically Based Rendering: To achieve realistic material effects, we employed programmatic node shaders for the main metallic surfaces. By layering noise textures and color gradient nodes, we precisely simulated the weathered patina of bronze and the high-gloss reflections of brass. For the luminous futuristic core, we utilized self-illuminating materials with postprocessing glow effects. All material parameters were grounded in physically based rendering principles, ensuring accurate rendering across various lighting conditions. 3. Immersive Space Review and Matching Verification: This is a critical step in the CAII workflow. The completed buckle model is imported into iTouch, along with a preset high-precision virtual leather bag model. Creators enter this virtual showroom via VR devices. Proportion and Comfort Evaluation: Through an immersive firstperson perspective, users can visually assess whether the buckle-tobag ratio is proportionate and whether the thickness and edge design cause friction that might be uncomfortable. Dynamic validation: A feature beyond the reach of traditional graphic design. Creators can virtually "pick up" a leather bag and manipulate the buckle's opening animation, observing in real-time the visual effects, structural dynamics, and leather deformation interactions during dynamic use. During this process, they discovered that the initially designed buckle pins were slightly too short, causing a visual mismatch with the leather buckle loop during the opening animation. This led to immediate revisions in the modeling phase. Multi-light source environment testing: The virtual engine enables one-click switching between various lighting scenarios (e.g., warm indoor lighting, bright daylight, or nighttime neon lights), allowing rapid evaluation of material and gloss performance under different conditions to ensure authentic design solutions. Figure 2 Virtual Engine Compatibility Effects Note: The image is sourced from the iTouch 2.3 Final Results and Process Reflection The final deliverables include: a 4K rendering of the "Ring of Time and Space" buckle, a dynamic video demonstrating its opening and closing mechanism, and an interactive experience video showcasing engagement with a virtual leather bag in a VR environment. This initiative successfully validates the immense value of the CAII Research Paper This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. DOI: 10.5281/zenodo.17852678 Page 17 workflow in accessory design: full-process digitization—from conceptualization to interactive prototypes—completed entirely in virtual environments, eliminating the need for physical prototyping and saving substantial time and costs. Dynamic validation loop: VR immersive experiences deliver critical design feedback, validating usability beyond visual appeal—a gamechanging advancement over traditional design processes. Efficient iterative design: Structural issues identified in virtual environments can be promptly returned to upstream stages for modification, forming an efficient closed-loop optimization process. Reflection: The challenge lies in the high demands for creators 'proficiency in 3D software operations and engine knowledge, coupled with the need for manual conversion between AI-generated concept designs and producible 3D models. Nevertheless, the workflow's forward-thinking approach and efficiency demonstrate significant potential for digital design in fashion categories such as accessories and footwear. This study demonstrates that the AI-driven collaborative design workflow powered by virtual simulation has revolutionized accessory design paradigms. It bridges the gap between static renderings and dynamic experience prototypes, resolves core challenges like accessory-vehicle compatibility and functional structure through VR immersion validation, and enables efficient style exploration via AI. Practical applications confirm this workflow significantly enhances design efficiency and feasibility, establishing a closed-loop "design-validation-iteration" cycle. While current challenges include technical bottlenecks in 2D-to-3D model conversion and elevated designer competency requirements, the maturation of AI-generated 3D models and interactive technologies positions this approach for broad applications in personalized customization, virtual presentations, and design education. This represents a pivotal advancement in the digital transformation of fashion design. References 1. Liu Naikang. The Impact of the AI Era on the Employment of Vocational Art and Design Students and Response Strategies[J]. Employment and Security, 2024, (07): 20-22. 2. Yang Yijing. Building an Inclusive Digital Economy Model to Promote High-Quality Development of China's Economy—China Academy of Information and Communications Technology Releases "China Digital Economy Development White Paper (2020)"[J]. China Science and Technology Industry, 2020, (08): 5-7. DOI:10.16277/j.cnki.cn11-2502/n.2020.08.003. 3. Zhang Jin, Gong Xiaoli, Li Changli, et al. Exploration and Practice of AIGC Empowerment in Virtual Simulation Experiment Teaching Projects[J]. Laboratory Science, 2024, 27(06): 29-32. 4. Gong Xinyue, Bai Dan. Content Restructuring, Method Innovation, and Teaching Reform Strategies for Basic Digital Painting Courses in the AIGC Era[J]. Art Education Research, 2025, (21): 128-130. 5. Li Jiaqi. Breaking Path Dependence: The Upgrade Path of China's Virtual Production Technology[J]. AudioVisual, 2024, (05): 29-32. DOI:10.19395/j.cnki.1674246x.2024.05.036. 6. Zuo Hanshuo, Wang Yixiong. Research on the Application of AIGC-Empowered Virtual Simulation Technology in Digital Media Art[J]. Toy World, 2025, (02): 174-176. 7. Huang Shenze, Lü Xin. Research on High-Fidelity Virtual Human Design Driven by AIGC Technology[J]. Art & Design Research (Chinese and English), 2024, (06): 95-101.