Measuring open innovation practices with ChatGPT and their effects on firm financial performance

This study proposes a fully automated, interpretable two-stage method using ChatGPT to measure open innovation practices (OIPs) from 10-K filings of Russell 3000 firms, addressing limitations in existing OIP measurement approaches, and examines how these practices influence firm financial performance. The method identifies eight nuanced OIP types, including novel dimensions like Supplier and Manufacturing Collaboration. OLS regression results show overall open innovation has no direct effect on firm financial performance, but specific practices (e.g. Technology and Platform Ecosystems) and R&D intensity jointly drive performance gains. Curvilinear analyses reveal no ‘paradox of openness’; instead, Supplier and Manufacturing Collaboration, Technology and Platform Ecosystems, and Marketing and Commercialization Cooperation exhibit U-shaped relationships, with performance improvement after an engagement threshold.