Sex and Gender in the Computer Graphics Research Literature

Current practices in computer graphics research mishandle sex and gender in ways that introduce algorithmic bias, with the paper arguing these established practices are "scientifically incorrect" and can produce harmful real-world effects in graphics systems and applications.

The problem is widespread enough to warrant a full literature survey, indicating that gender and sex treatment is not an isolated issue but rather a systematic concern across the computer graphics research community that requires collective attention.

These issues can be addressed as technical limitations, suggesting that researchers have concrete, implementable solutions available—the bias is not an inevitable feature of graphics research but rather a solvable engineering problem.

Algorithmic fairness frameworks apply to computer graphics, extending responsible AI concerns beyond traditional machine learning domains into visual computing, meaning graphics researchers should adopt fairness-focused practices similar to those in other AI fields.

The paper identifies a gap between scientific understanding and research practice, implying that the computer graphics community may not be applying current scientific knowledge about sex and gender correctly, pointing to a need for better education and standards in the field.