[RESEARCH PAPER] Measuring the Openness of AI Foundation Models, by Thibault Schrepel and Jason Potts

Measuring the Openness of AI Foundation Models: Competition and Policy Implications

This paper provides the first comprehensive evaluation of AI foundation model licenses as drivers of innovation commons. 

Thibault Schrepel and Jason Potts introduce their analysis by outlining how AI licenses regulate access privileges to the fundamental inputs of AI innovation commons.  They show that AI licenses operate as a bottleneck, as their level of openness directly influences the flow of knowledge and information into the commons. 

Then they introduce a new methodology for evaluating the openness of AI foundation models. Their methodology extends beyond purely technical considerations to more accurately reflect AI licenses’ contribution to innovation commons. They proceed to apply it to today’s most prominent models—including OpenAI’s GPT-4, Meta’s Llama 3, Google’s Gemini, Mistral’s 8x7B, and MidJourney’s V6—and find significant differences from existing AI openness rankings. 

They conclude by proposing concrete policy recommendations for regulatory and competition agencies interested in fostering AI commons based on their findings.

.

About the authors:

Thibault Schrepel is Associate Professor of Law at the Vrije Universiteit Amsterdam, and Faculty Affiliate at Stanford University CodeX Center.

Jason Potts is a Distinguished Professor of Economics at RMIT University, and Co-director of the Blockchain Innovation Hub at RMIT.