Accelerated Knowledge Discovery: A Vision for NASA Science

This paper introduces the sixth paradigm of scientific discovery: accelerated knowledge discovery (AKD). This paradigm is defined by the full integration of artificial intelligence (AI) into the research workflow as a tool augmenting cognitive capabilities of human scientists. AKD emerges from the convergence of advanced AI models, autonomous agentic systems, and human‐AI collaboration. AKD accelerates the research cycle by reducing the time from conceptualization to discovery. It automates labor‐intensive tasks such as literature review, hypothesis generation, experimental design, data analysis, modeling, simulation, and manuscript drafting. In well‐defined domains, AKD can transform the scientific method into a continuously adaptive cycle, where outputs from each phase inform the next. These closed‐loop scientific workflows shorten discovery timelines and reduce overhead. In addition to the scientific speed up, AKD targets an increase in the quality of research allowing for more systematic discovery of knowledge. However, AKD's success depends on principled, trustworthy design. This requires a holistic approach that emphasizes explainability, reproducibility, robustness, adaptability, and transparency. Key requirements include alignment with open science principles, required human oversight, scientific accountability, and rigorous provenance tracking. Human researchers must remain ultimately responsible for scientific integrity, ethical reasoning, and interpretation, with AI serving as an augmentative partner. Although the proposed approach applies to any domain focused on scientific discovery, this paper highlights AKD's potential to advance NASA's science mission, given the agency's vast data assets, complex objectives, and interdisciplinary challenges. By integrating NASA's data, foundation models, scalable computing, and knowledge frameworks, AKD can accelerate discovery and foster innovation across its scientific portfolio.