| Abstract: |
This study analyzes the current state of artificial intelligence (AI)
technologies for addressing and mitigating climate change in the manufacturing
sector and provides an outlook on future developments. The research is
grounded in the concept of general-purpose technologies (GPTs), motivated by a
still limited understanding of innovation patterns for this application
context. To this end, we focus on global patenting activity between 2011 and
2023 (5, 919 granted patents classified for “mitigation or adaptation against
climate change” in the “production or processing of goods”). We examined time
trends, applicant characteristics, and underlying technologies. A topic
modeling analysis was performed to identify emerging themes from the
unstructured textual data of the patent abstracts. This allowed the
identification of six AI application domains. For each of them, we built a
network analysis and ran growth trend and forecasting models. Our results show
that patenting activities are mostly oriented toward improving the efficiency
and reliability of manufacturing processes in five out of six identified
domains (“predictive analytics”, “material sorting”, “defect detection”,
“advanced robotics”, and “scheduling”). Instead, AI within the “resource
optimization” domain relates to energy management, showing an interplay with
other climate-related technologies. Our results also highlight interdependent
innovations peculiar to each domain around core AI technologies. Forecasts
show that the more specific technologies are within domains, the longer it
will take for them to mature. From a practical standpoint, the study sheds
light on the role of AI within the broader cleantech innovation landscape and
urges policymakers to consider synergies. Managers can find information to
define technology portfolios and alliances considering technological
co-evolution. |