Detail from a patent for a toy that might look familiar (Image: Public domain)

Patents are one of the best sources of data on technology development — an open-ended, historical and adaptive system that shows us how and why inventions have come to be. But is the U.S. patent system broken?

That question is being raised more frequently these days, as inventors and companies operate in an increasingly competitive ecosystem. For Jose Lobo and Deborah Strumsky, both Fellows of the ASU-SFI Center for Biosocial Complex Systems, it’s a question that deserves careful consideration.

Bringing together experts from academia, industry and the legal profession, Lobo and Strumsky are hosting an Applied Complexity Network (ACtioN) working group at SFI March 12-14 to explore the nature of the patent system. And with a complex system that has evolved over the course of 224 years — from hundreds of technologies to hundreds of thousands — there’s much territory to explore.

“Part of what makes SFI’s approach unique, and what’s makes this working group unique, is that SFI researchers want to go beyond talking about analyzing the streams of output from the patent office,” Strumsky explains. “ is working group calls for a much deeper look into production of information and how it affects our understanding of our world.”

The patent system’s detailed and precise descriptions of inventions, including data on the inventor, where they worked and when they were working, o ers researchers a way to understand technological change in terms of selection, obsolescence, adaptation, and diffusion processes. But the system is also human-powered, and humans are awed. Can AI help?

Some countries are already using AI to help patent examiners. e strategy seems to be working. AI has been able to replace redundant searches, freeing patent examiners to work with clients and be more responsive to their needs. Examiners can get overwhelmed, but AI never gets tired.

“We need to understand the nature of the patenting system as an information processing and generating system before we can assess how AI can make the patenting system better,” Lobo says.

“More generally, before a diagnosis of ‘the system is broken’ is meaningfully made, and a solution proposed, we need to understand the fundamental nature of the system.”

To understand that fundamental nature, Strumsky says researchers need to study the patent system as an evolving system that generates enormous amounts new information on a daily basis. “We need to understand how the type of information generated enables and constrains our ability to study technology and understand how we interact with it.” 

Read more about the working group.