How Generative AI can be a key enabler for Smart Additive Manufacturing

Generative AI is emerging as a transformative approach in the manufacturing domain, with significant potential to enhance in-situ monitoring and in-line qualification in Additive Manufacturing (AM). This presentation provides an overview of the current capabilities, limitations, and application potential of generative models, particularly Generative Adversarial Networks (GANs) and diffusion models, for addressing data scarcity and imbalance in industrial AM scenarios. While traditional AI systems often rely on large volumes of well-balanced data, the rarity and high cost of capturing real-world anomalies and defects in AM processes pose a critical challenge. Generative models not only offer a promising solution by enabling the synthetic creation of representative process data, but they also open new pathways towards zero-defect manufacturing in real production scenarios across several industrial sectors.