OpenAI Foundation will fund Ruxandra Teslo’s proposal for using “biotech’s lost archive” — detailed materials from failed companies, including regulatory filings and safety data — to train medical AI systems through its new Data for Public Health effort. The foundation said early data grants total $40 million for novel cancer vaccine dataset work at the University of North Carolina, Chapel Hill, and support for OpenAdmet competitions to predict drug effects. Teslo’s initial request was funded as well through a $500,000 grant, with the work to be pursued by 1Day Sooner, an advocacy group representing clinical trial volunteers. The move signals growing pressure for biology-grade, high-quality datasets as a constraint on reliable AI performance in drug development and regulatory analysis.