Research Ethics and Responsible Computation

Most computational mathematics does not involve human participants, but authors must obtain appropriate ethics review and consent when datasets, experiments or applications involve identifiable people, sensitive records or interventions.

Computational work should report limitations, uncertainty and failure modes honestly. Authors must not fabricate numerical evidence, conceal unstable calculations, misrepresent benchmark comparisons or release code and data that violate privacy, security, licences or contractual obligations.