Algorithmic Justice League — publication: Gender Shades study (2018) by Buolamwini and Gebru demonstrated intersectional accuracy disparities in commercial facial recognition from IBM, Microsoft, and Face++. Dark-skinned women had highest error rates; light-skinned men had lowest. Over 4,900 citations on Semantic Scholar.

Note: The source URL points to a Semantic Scholar page for the Gender Shades study, but the provided source text contains only a JavaScript error message stating the page requires JavaScript to load. No actual content about the study, citation count, authors, findings, or any other claim details is accessible from the provided source text. While the URL structure suggests it is the correct paper (matching authors Buolamwini and Gebru, and the paper title), the actual page content cannot be verified from this excerpt. The citation count (4,900+ as of 2026-03) cannot be confirmed or contradicted because the source does not display any data. This is a technical access issue, not a contradiction of the claim itself.

Note: The provided source text is a JavaScript verification page, not the actual Semantic Scholar paper content. It contains no information about the Gender Shades study, the researchers (Buolamwini and Gebru), the findings regarding intersectional accuracy disparities, the specific error rates for different demographic groups, or the citation count. While the URL appears to point to the correct paper, the actual source content provided does not address any aspect of the claim. Without access to the actual paper or metadata, the claim cannot be verified against the source.