‘Advancing the public interest’ is a criterion for de-identified data use for research via several national data platforms and biobanks. This may be referred to via cognate terms such as public benefit, public good or social value. The criterion is often adopted without it being a legal requirement. It is a legal requirement in some jurisdictions for sharing identifiable data without consent, which does not apply to de-identified data. We argue that, even in circumstances where there are few or no legal restrictions on the sharing of de-identified data, there is a sound ethical reason for platforms to nevertheless impose a public interest criterion on data sharing. We argue that a public interest test is ethically essential for justifying research use of de-identified data via government-funded platforms because (1) it allows to promote public good and to minimise potential harmful consequences of research for both individuals and groups, for example, by offering grounds to reject research that could lead to stigmatisation of marginalised populations; (2) national data platforms hold public data and are made possible by government funds, and therefore should be used to support public interests and (3) it can demonstrate trustworthiness and contribute to promoting the social licence for data platforms to operate, which is especially important for efforts to align data governance policies with public norms and expectations.