Policy Pulse: Bridging the Gap Between Scientific Research and Public Accessibility Through Generative AI and Summarization Models

The complexity and inaccessibility of scientific research papers often create a significant barrier between the scientific community and the general public, hindering effective communication and delaying informed policymaking. Policy Pulse addresses this gap by utilizing a Retrieval-Augmented Generation (RAG) model to transform complex research papers into concise, comprehensible policy briefs. The model processes research documents through three main stages: preprocessing, retrieval, and generation, leveraging technologies like vector embeddings and large language models to produce high-quality summaries. Experiments demonstrated that Policy Pulse captures core semantic insights effectively, achieving high BERT scores compared to alternatives like OpenAI and Google Gemini, despite lower ROUGE-L scores indicating room for improvement in surface-level precision. The solution prioritizes intellectual property protection, user customization, and accessibility by maintaining a reading level suitable for non-experts. Challenges include processing efficiency, ethical considerations in usage, and limitations in multilingual capabilities. Future development aims to enhance interactivity, address feedback integration, and expand multilingual support to improve global applicability. Policy Pulse presents a novel approach to bridging the divide between science and society, fostering informed policymaking and public engagement.