Bethlehem Steel in Bethlehem, Pennsylvania was one of the world's leading steel manufacturers for most of the 20th century until 1982, when it discontinued most of its operations. Credit: J. Schnalzer
A new study suggests AI-powered technologies may play a similar role in society as industrial research in the early 20th century by reducing coordination costs and potentially reshaping how collective invention is organized.
To show the shift from an agricultural economy to science and innovation around the 20th century, the researchers digitized roughly half a million pages of historical patent yearbooks covering 1.6 million patents and linked these to census records and surveys of industrial research laboratories. This dataset allowed them to track inventors’ occupations, collaborations, organizational affiliations and the technological novelty of patents.
The findings, published in the journal Research Policy, show how invention reconcentrated in large cities, especially on the East Coast and in the Rust Belt, and the new innovation system exhibited drastically lower participation rates among women and foreign-born inventors.
At the level of inventors, the data shows lengthening learning curves, a rapid rise of engineers, a greater reliance on academic literature, and the emergence of academic patenting. At the meso-level, there is a sudden shift to teamwork that is no longer coordinated through family ties but through organizational ties within firms and labs.
“Over the last decade, we have seen a revival of R&D labs driven by tech giants, like Google, Facebook, and Amazon,” said Complexity Science Hub’s Frank Neffke. “As in much of the 20th century, when behemoths like Bell Labs not only patented, but also pushed the scientific frontier, spawning several Nobel prize winners, important breakthroughs in AI are nowadays driven by industrial, not academic labs.”
The researchers propose future examination of whether today’s organized corporate research once again promotes radical versus incremental innovation, to what extent women and immigrants participate, and how it may change the geographic concentration of innovation.
Data from Harvard Kennedy School