Nature misses a chance to grant agency to marginalized communities in inviting Alice Popejoy to point out that altering racial classifications will not absolve power imbalances in genetics (Nature 596, 463; 2021).

In 1785, philosopher Christoph Meiners reduced continental-scale diversity to an imperial classification system to subjugate colonized peoples. This system is still used by geneticists, and lingers beyond terms such as Caucasian. New ethnonyms replaced older terms (‘mongoloid’ became ‘Asian’, for instance) but failed to redress underlying racism. And socially constructed categories are used in biologically essentialist ‘race correction’ to model disease risks (see, for example, D. E. Roberts Lancet 397, 17–18; 2021).

To demolish genetic racism, geneticists must defer to communities to self-define their ‘belongingness’ (see, for example, K. S. Tsosie Curr. Opin. Genet. Dev. 62, 91–96; 2020). Any unequal system of classification that reifies race, ethnicity and ancestry for biological insight reproduces the obstacles it attempts to dismantle and does not solve the causes of health disparities.

We advocate empowering communities to label themselves; to undertake ethnographies to contextualize research findings; and to self-determine research they deem beneficial.