The "Gold Rush" in AI and Robotics Patenting Activity. Do innovation systems have a role?

This paper studies patenting trends in artificial intelligence (AI) and robotics from 1980 to 2019. We introduce a novel distinction between traditional robotics and robotics embedding AI functionalities. Using patent data and a time-series econometric approach, we examine whether these domains share common long-run dynamics and how their trajectories differ across major innovation systems. Three main findings emerge. First, patenting activity in core AI, traditional robots, and AI-enhanced robots follows distinct trajectories, with AI-enhanced robotics accelerating sharply from the early 2010s. Second, structural breaks occur predominantly after 2010, indicating an acceleration in the technological dynamics associated with AI diffusion. Third, long-run relationships between AI and robotics vary systematically across countries: China exhibits strong integration between core AI and AI-enhanced robots, alongside a substantial contribution from universities and the public sector, whereas the United States displays a more market-oriented patenting structure and weaker integration between AI and robots. Europe, Japan, and South Korea show intermediate patterns.

💡 Research Summary

The paper investigates patenting trends in artificial intelligence (AI) and robotics from 1980 to 2019, introducing a novel classification that separates traditional robotics from robotics that embed AI functionalities. Using a comprehensive patent‑family dataset drawn from PATSTAT, the authors combine Cooperative Patent Classification (CPC) codes, keyword filters, and document‑level text mining to construct three mutually exclusive technological domains: core AI, traditional robots, and AI‑enhanced robots. Core AI captures foundational AI capabilities such as machine learning, reasoning, perception, and autonomous decision‑making. Traditional robots encompass industrial, service, and social robots as defined in prior literature. AI‑enhanced robots are those in which AI is structurally integrated, providing learning, perception, or adaptive control.

The empirical strategy proceeds in three stages. First, unit‑root tests (ADF, PP) confirm that the patent counts for each domain are non‑stationary, justifying a time‑series approach. Second, structural break analysis using Zivot‑Andrews and Bai‑Perron procedures identifies a concentration of breakpoints after 2010, indicating a regime shift coinciding with the diffusion of AI capabilities into robotics. Third, the authors estimate pairwise cointegration relationships across the three domains for a set of major innovation systems: China, the United States, the European Union (represented by leading EU members), and Japan/Korea.

Key findings are as follows. (1) Patent activity in AI‑enhanced robots accelerates sharply from the early 2010s, outpacing both core AI and traditional robotics, which follow more gradual trajectories. (2) Structural breaks are clustered in the post‑2010 period, supporting the notion of an “AI gold rush” that reshapes the dynamics of robotic innovation. (3) Long‑run relationships differ systematically across countries. In China, core AI and AI‑enhanced robots exhibit strong cointegration, and universities and public‑sector entities contribute a substantial share of patents (over 30 %). This reflects a state‑coordinated innovation system that actively steers AI‑robot convergence. In the United States, the patenting structure is market‑driven; integration between AI and robotics is weaker at the applicant level, suggesting that private firms pursue AI and robotics largely in parallel rather than as a tightly coupled ecosystem. Europe shows intermediate patterns, with a mix of public‑policy initiatives and market forces leading to moderate integration. Japan and South Korea display similar intermediate dynamics, maintaining strong traditional robotics bases while gradually incorporating AI functionalities.

Policy implications derived from the analysis emphasize the need for tailored support mechanisms that reflect national innovation system characteristics. For countries like China, reinforcing university‑industry‑government collaboration can sustain the strong AI‑robot integration. For market‑oriented economies such as the United States, incentives that promote cross‑domain collaboration (e.g., joint R&D programs, standards development) may enhance the synergies between AI and robotics. The authors also stress the importance of early standard‑setting, ethical guidelines, and safety regulations to mitigate potential negative externalities of rapid AI‑robot diffusion.

Methodologically, the study contributes by demonstrating how detailed patent classification and advanced text‑mining can overcome the fuzzy boundaries between AI and robotics, enabling a more granular view of technological co‑evolution. The combination of structural break detection and cointegration analysis provides a robust framework for assessing whether emerging technologies share common stochastic trends or evolve independently within different institutional contexts.

Overall, the paper offers a comprehensive, data‑driven portrait of the “gold rush” in AI and robotics patenting, highlighting how national innovation systems shape the pace and nature of AI‑enhanced robotic development. Its findings are relevant for scholars of technological change, policymakers designing AI strategies, and industry leaders seeking to navigate the rapidly converging AI‑robot landscape.