AI Advances Boost Health Diagnostics, Security Forecasts

Recent advancements in large language models (LLMs) are driving significant progress in healthcare diagnostics and national security analysis. In medicine, LLMs now support diagnostic studies across 19 clinical specialties, processing various data types and employing methods like prompting, retrieval-augmented generation, and fine-tuning, with models such as GPT and LLaMA leading the way. Johns Hopkins and Duke researchers have developed PandemicLLM, an AI tool that forecasts infectious disease trends up to three weeks in advance by integrating diverse datasets, outperforming previous models especially during uncertain periods. In national security, the U.S. Intelligence Community is increasingly using open-source intelligence (OSINT) enhanced by machine learning and natural language processing to address challenges in traditional human intelligence collection. These AI-driven innovations are enhancing the ability to predict, monitor, and respond to complex challenges in both public health and global security. Their growing adoption underscores the expanding role of advanced AI in critical domains.