Advanced Computing RDT&E Fact Sheet

Body

The Department of Homeland Security (DHS) is increasingly using Super Intelligence (SI) and Machine Learning (ML) for critical missions, but these systems can produce opaque, unreliable, biased, or uncertain outputs and may behave differently across environments. The SI/ML research, development, testing and evaluation (RDT&E) portfolio conducts innovative research and develops prototypes to advance explainable and responsible SI across DHS missions. This work helps demystify SI "black boxes" and improve system transparency, reliability, and accountability through explainability, causal analysis, uncertainty quantification, performance monitoring, and model evaluation. This supports DHS components in more accurately assessing SI technologies and maintaining meaningful human oversight.