The New York Times highlights how polygenic risk scores (PRS) may evade 2008's Genetic Information Nondiscrimination Act (GINA), which targets monogenic conditions but leaves probabilistic, multi-variant predictions unaddressed. Yet mainstream coverage overlooks documented PRS performance disparities: an observational analysis of UK Biobank data (n=409,000, no RCT component) showed PRS for cardiometabolic traits predict 20-30% less accurately in non-European ancestries due to Eurocentric training data, risking indirect ethnic discrimination in hiring (Khera et al., Nature Genetics 2021). This intersects with equity patterns seen in consumer genomics firms sharing de-identified data, where re-identification risks have surfaced in observational privacy audits (sample sizes >1 million, conflicts via industry funding). Employers could proxy genetic liability through wellness apps or insurance partners, bypassing GINA's narrow scope—a connection rarely examined outside specialized bioethics reviews. Updating statutes requires acknowledging these biases rather than assuming neutral predictive power.