Build Verifiable Apps Using Automated Reasoning and Generative AI

Explore methodologies for combining large language models with automated reasoning tools to create more reliable AI systems in this 42-minute conference talk from AWS re:Inforce 2025. Discover how formal methods can guide and constrain generative AI while maintaining creative capabilities and ensuring verifiable outputs. Learn about the complementary strengths of probabilistic and symbolic approaches through practical demonstrations of hybrid systems that merge these technologies. Examine real-world applications including how Amazon Q Developer and Amazon Bedrock Guardrails leverage automated reasoning to generate safe, logically correct output free from hallucinations. Gain insights into building AI applications that balance creativity with rigorous verification through the integration of generative AI and formal verification methods.