A governance and ethics framework for computational reproductive selection modeling

Computational models of reproductive selection occupy an unusual ethical position: they are not clinical interventions, yet they formalize exactly the kind of reasoning — choosing which traits to favor across generations — that bioethics has scrutinized for decades in the context of selective reproduction. This paper proposes an explicit governance and ethics framework for such modeling work, developed alongside the CORGData governed synthetic pool. The framework rests on five principles: auditability of every modeling choice, a strict prohibition on person-level reconstruction, transparency of policy parameters, restriction of long-range applications, and reversibility of published claims through versioned pre-registration. It also provides an operational checklist that any simulation study in this space should satisfy before publication. The argument is normative and analytical rather than empirical. No clinical, eugenic, or population-policy recommendation is made or implied; the framework governs models, not people.