Investigate lightweight masking algorithms such as format-preserving encryption, to preserve data structure while obscuring sensitive values [103]. This ensures compatibility with real-time AI applications running on [[Substrate]]'s high-performance runtime. Format-preserving encryption allows masked data to retain its original format (e.g., a masked credit card number still looks like a valid number), enabling seamless integration with existing AI pipelines that rely on structured data, thus avoiding the need for extensive data reformatting while maintaining compatibility with Substrate's weight-based execution model. See also: [[ZKP/Research/Masking/Integration with ZKP|Integration with ZKP]]