This paper looks at whether scientific collaborations across Chinese regions help local firms file more artificial intelligence (AI) patents. The authors focus on co-authored research papers that link two different provincial areas. They ask if those “co-publication” ties act as knowledge spillovers that raise regional AI patenting, and whether the mix of organizations doing research in a region changes that effect.

To study this, the researchers combine publication and patent records for 26 Chinese provincial-level regions over the years 2006–2020. They use a network-econometric approach, which means they treat co-authorship links between regions as a network and then use statistical models to see how those links relate to later patenting. This shifts attention from the usual measures of proximity or inventor partnerships to the scientific layer carried by co-authored papers.

The main finding is that interregional co-publication spillovers do foster AI patenting. In other words, regions that are connected to other places through joint AI research tend to produce more AI patents. But this benefit is not evenly shared. The study finds that the patent gains concentrate almost entirely in regions where privately owned enterprises dominate research and development (R&D). Regions where public institutions or state-owned actors lead R&D do not capture the same patent payoff from scientific ties.

Why this matters: co-authored research can do more than advance basic knowledge. It can move ideas across regions and help convert them into patented applications. The paper suggests that the structure of local R&D — especially the presence of private firms that can absorb and apply new knowledge — determines whether a region turns scientific connections into technological output.