The genomics field, where bioinformatics algorithms play a significant role, remains deeply entangled with societal inequalities.
A good example of how racism and technology combine to perpetuate healthcare inequality comes from a paper by researchers at Beihang University. They quantified how racism corrupts both cancer gene discovery and patient survival predictions using The Cancer Genome Atlas (TCGA), one of the world’s largest cancer genomics databases.
The paper’s findings are infuriating; with white patients comprising over 75% of the cancer dataset, AI models identified cancer genes that have ZERO overlap with those affecting Black and Brown patients. So essentially, the “breakthroughs” celebrated in precision medicine are breakthroughs for white bodies only, effectively codifying white supremacy at the molecular level.
As has been stated many times, this is not a bug; it is a feature of a system designed by and for whiteness. TCGA, despite its million-dollar budget and promises of revolutionising cancer care, consists of data that overwhelmingly benefits white patients. The researchers also found that using race as a “confounding factor” could not fix the fundamental problem. Showing there is no algorithmic adjustment out of structural racism further underscores why technosolutionist approaches will never be enough.
A truly sustainable solution for this requires that we understand and fully acknowledge how medical racism already exists outside the algorithm and then radically change the systems that uphold it.
See: Racial Bias Can Confuse AI For Genomic Studies at Tech Science Press.
Photo by Ekke Krosing on Unsplash.