When every hypothesis is published: AI and the exploitation of Open Science Assets

This presentation was given at a 27 October 2026 event hosted by NBIS – SciLifeLab. The aim of the presentation was to highlight how open data are being analysed in the GenAI era, with sharp increases in publications analysing large epidemiological, nutrition and health datasets, including NHANES, CDC WONDER and TriNetX. This is producing fragmented outputs that describe the data, rather than comprehensive works that analyse the data. The presentation also demonstrates the methods by which these mass-produced outputs are generated, and possible solutions for curation in a world where every data-driven hypothesis can be rapidly tested and reported in the scientific literature.