In recent years, additive manufacturing (AM) of metallic systems has become prevalent for manufacturing and/or repairing a wide range of industrial components. However, the AM industry is still grappling with issues related to inconsistent quality and properties of the AM built parts. These issues are major impediments to meeting the desired manufacturing and performance standards. Digital twins (DT) are being adopted in the AM industry to optimize the entire manufacturing process and enable products with high quality and performance attributes to be produced in a repeatable manner. Current methods for using AI/ML approaches for AM DT often assume a direct correlation between AM input parameters and the relevant output properties. These approaches often ignore material characterization information and/or proper information-rich AI surrogate models (thereby having no capability for uncertainty quantification (UQ) prediction needed for a credible AM digital twin framework). This opportunity is for developing a suitable DT framework that couples materials-based AM surrogate models with rigorous physics-based models for AM, and for identifying suitable approaches for integrating these surrogate models into an AI/UQ framework. This capability is expected to provide a detailed understanding of the structure/property relationships for calibrating an AM DT framework for suitable AM processes such as directed energy deposition (DED) and laser powder bed fusion (L-PBF).