Abstract
Artificial intelligence is transforming the sciences, yet generalconversational AI systems often generate unverified "hallucinations"undermining scientific rigor. We present OceanAI, a conversational platformthat integrates the natural-language fluency of open-source large languagemodels (LLMs) with real-time, parameterized access to authoritativeoceanographic data streams hosted by the National Oceanic and AtmosphericAdministration (NOAA). Each query such as "What was Boston Harbor's highestwater level in 2024?" triggers real-time API calls that identify, parse, andsynthesize relevant datasets into reproducible natural-language responses anddata visualizations. In a blind comparison with three widely used AIchat-interface products, only OceanAI produced NOAA-sourced values withoriginal data references; others either declined to answer or providedunsupported results. Designed for extensibility, OceanAI connects to multipleNOAA data products and variables, supporting applications in marine hazardforecasting, ecosystem assessment, and water-quality monitoring. By groundingoutputs and verifiable observations, OceanAI advances transparency,reproducibility, and trust, offering a scalable framework for AI-enableddecision support within the oceans. A public demonstration is available athttps://oceanai.ai4ocean.xyz.