Shanghai Made a Clean Dataset Count as a Publication

The rule, in force from 1 November, puts high-quality datasets into research evaluation and explicitly covers unpublished and negative results.

Shanghai’s science and technology commission and data bureau jointly issued the Shanghai Science Data Management Measures on 13 November 2026, the city’s first dedicated government rule on the subject, comprising seven chapters and 28 articles and effective from 1 November 2026 through 31 October 2028. The central change is attribution: high-quality science datasets now count as research output in the evaluation system, with results feeding into institutional performance reviews and project applications, and unpublished and negative data explicitly brought into scope. The rule also requires submitting data before performance evaluation, allows project budgets to cover data governance, and authorises market participants to take part in cleaning, annotation and value-added work through a whitelist mechanism. The commercial read is that AI4S has been blocked less by models than by a supply of clean, citable data, and this converts data work from an unrewarded chore into a scored contribution.