Paper Type
ERF
Abstract
A growing body of evidence shows that AI innovations are more process-oriented and less radical than comparable IT innovations. Conventional innovation theory predicts less radical innovations should yield lower economic returns. Yet, we document an AI appropriability paradox: AI patents command a significant value premium over matched IT patents despite lower radicalness. Applying Blinder–Oaxaca decomposition from labor economics, method rarely used in IS research, to matched patents spanning twenty years, we find only about one-third of this gap is attributable to observable patent characteristics. The remaining unexplained premium cannot be accounted for by standard Schumpeterian dimensions. Drawing on appropriability regime theory and general purpose technology economics, we interpret this premium as reflecting a structurally different appropriability regime governing AI innovations where the sources of value extend beyond technological radicalness. This study challenges the assumption that radicalness drives value and reframes how scholars and managers evaluate AI innovation portfolios.
Paper Number
1927
Recommended Citation
Mahmud, Jishan, "Creative Destruction or Engines of Innovation? A Comparative Analysis of AI and IT Innovation Using Stock Market Reactions" (2026). AMCIS 2026 Proceedings. 5.
https://aisel.aisnet.org/amcis2026/dite/sig_dite/5
Creative Destruction or Engines of Innovation? A Comparative Analysis of AI and IT Innovation Using Stock Market Reactions
A growing body of evidence shows that AI innovations are more process-oriented and less radical than comparable IT innovations. Conventional innovation theory predicts less radical innovations should yield lower economic returns. Yet, we document an AI appropriability paradox: AI patents command a significant value premium over matched IT patents despite lower radicalness. Applying Blinder–Oaxaca decomposition from labor economics, method rarely used in IS research, to matched patents spanning twenty years, we find only about one-third of this gap is attributable to observable patent characteristics. The remaining unexplained premium cannot be accounted for by standard Schumpeterian dimensions. Drawing on appropriability regime theory and general purpose technology economics, we interpret this premium as reflecting a structurally different appropriability regime governing AI innovations where the sources of value extend beyond technological radicalness. This study challenges the assumption that radicalness drives value and reframes how scholars and managers evaluate AI innovation portfolios.
Comments
SIG DITE