Artificial Intelligence and Punctuated Equilibrium

How do government attention and policy agendas shift when information constraints are relaxed through technological advancement? Punctuated equilibrium theory links agenda-setting dynamics to governments’ limited information-processing capabilities, yet we know little about whether and how technology improvement such as artificial intelligence (AI) influences policy agendas. This article proposes a conditional digital punctuation framework to specify the conditions under which technological advancements drive agenda shifts. Using China’s implementation of AI initiatives as a quasi-natural experiment, we find that government attention exhibits a punctuated pattern. AI increases technology-focused attention, signaling incremental changes, while having minimal impact on general government attention. This selective impact stems from AI’s asymmetric effects on information processing. Specifically, AI significantly increases the procurement-based indicator of information quantity, which is associated with greater attention to technology sectors; however, no significant change is observed in the corresponding indicator of information quality, consistent with its limited impact on general government attention.

Room : Sciences po, room G009, 28 rue des Saints-Pères, 75007, Paris

Speaker: Li Yiran (Hong Kong Baptist University)

Discussant: Stéphanie Balme (CERI-Sciences Po/CNRS)

Chair: Jérôme Doyon (CERI-Sciences Po/CNRS)

Scientific coordinator : Jérôme Doyon (CERI-Sciences Po/CNRS)