JMIR Publications

ABSTRACT

Frontier AI developers increasingly state that they are training models to promote human flourishing, yet the term is rarely defined, operationalized, or evaluated in public. This editorial argues that flourishing should become a third axis of model evaluation and governance, alongside capability and safety, and that the minimal step toward it is transparency rather than consensus on values. We propose a Human Flourishing Statement: a concise, published disclosure of the normative target of alignment—its working definition, the framework informing it, the dimensions included and excluded, how trade-offs and disagreements are handled, and how developers evaluate whether training moves models toward those goals. Benchmarks that score models on flourishing are now emerging, but a score is not a disclosure: it cannot reveal which conception a developer endorses. Drawing on health research's long tradition of making value-laden choices transparent, from reporting guidelines to ethics review, we argue that the health informatics community is well positioned to lead the development of reporting standards for flourishing-aware AI.