AI-Driven 3D Printing: Generative Gcode for Quality and Efficiency

Artificial Intelligence (AI) is transforming the landscape of three-dimensional (3D) printing by enabling generative approaches that extend beyond conventional, parameter-tuning strategies. Rather than merely adjusting nozzle temperatures or speeds, modern AI models can create or refine Gcode paths themselves, seeking to balance objectives such as print strength, dimensional accuracy, and defect minimization. This paper introduces a holistic AI-driven workflow for automated Gcode generation. We demonstrate that generative Large Language Models (LLMs) can parse, synthesize, and optimize printing instructions, thus offering an alternative to human-engineered slicing parameters. Our approach incorporates a continuous feedback loop in which a predictive model evaluates each generated Gcode script for potential print outcomes. Experimental results illustrate strong correlations between the number of lines in generated Gcode and key performance metrics (tensile strength, defect rate, and dimensional accuracy). Notably, some scripts with fewer lines reduced material defects yet slightly degraded dimensional accuracy, while more complex scripts improved measurement fidelity but also introduced minor flaws. These findings illuminate the trade-offs in AI-driven 3D printing and highlight the method's capacity for ongoing improvement. By exploring future expansions, such as continuous-function printing trajectories and multi-material prints, this study shows that AI-based Gcode generation can significantly advance 3D printing quality and customization across diverse applications.