news
February 11, 2026
lisa daigle
Deputy editor
military embedded systems
arlington, virginia The Defense Advanced Research Projects Agency (DARPA) has launched what it calls the Constructive Learning and Reasoning for AI Complex Systems Engineering (CLARA) program and solicited proposals for high-confidence artificial intelligence (AI) research.
According to DARPA’s initial solicitation, CLARA aims to create a theory-driven, algorithm-based, highly reusable, and scalable foundation for high-assurance/broadly applicable AI components used in a variety of defense and commercial sectors. The agency notes that the current industry approach to AI is to incorporate specialized automated reasoning (AR) components into large-scale language models (LLMs) or other similar machine learning (ML) systems. According to DARPA officials, these ML-centric systems typically have weak guarantees and lack real safeguards.
DARPA’s goal in the CLARA research program is to tightly integrate AR and ML components to create reliable AI, bringing together two distinct fields of AI to provide the speed and flexibility of ML with AR proof-based verifiability with strong logical explainability and computational tractability.
According to DARPA’s announcement, CLARA is expected to create a powerful method for hierarchical, fine-grained, and transparent composition of important types of ML and AR components, including Bayesian, neural net, and logic programs.
Featured companies
