Analyzing How AI Technology Evolves through Diffusion

Missing media item.

Artificial intelligence (AI) is arguably one of the most significant technological innovations as it reshapes the way we think about work, organizing and competitions. As AI's role in organizing and innovations continues to expand, we must understand how AI technology evolves as it continues to get diffused throughout different industries. Youngjin Yoo, xLab founder and the Elizabeth M. and William C. Treuhaft Professorship in Entrepreneurship, is working on a study funded by the National Science Foundation (NSF) that will investigate how AI emerges from complex and dynamic interactions across different technological components that are often independently designed and how AI innovations evolve as it traverses from one industry to another. 

“Through this research, we can explore how the meaning of AI innovation evolves as the boundary of AI's innovation ecology continues to shift and expand,” says Yoo. “Through a comprehensive historical analysis of contemporary AI innovation, we will better understand how AI has evolved and will continue to evolve in the future. This identifies potential weak links in the AI innovation ecology for continuing development of AI innovation to direct future investments and research efforts.”

The team is using a multi-disciplinary research method with two distinct but interrelated research activities. First, they are conducting a qualitative analysis of the AI innovation's emergent evolution, leveraging archival data about the development of current AI innovation from news articles about AI, related enabling technology components, and its applications. Second, they will conduct computational analyses to learn how to understand the generative diffusion of AI innovations over time through different fields by analyzing:

  • Publicly available documents, such as mainstream news, academic research publications, and
  • Open-source projects in the GitHub platform that use open-source AI frameworks.

Specifically, they are leveraging recent computational tools, namely the Relational Graph Convolutional Networks method, in studying the evolution of innovation ecology. These analyses aim to identify the dynamic patterns by which AI innovation is evolving and moving through.