Balancing openness and ownership: open innovation strategies for AI development

Purpose This paper explores how firms configure open innovation (OI) strategies when integrating artificial intelligence (AI) into their innovation models. Through the case of Baidu, it examines how OI contributes to business model innovation, highlighting how firms navigate the tension between openness and ownership in AI development. Design/methodology/approach Adopting an exploratory case study approach, the research employs big data analysis methods, including thematic network and collaboration cluster analyses. These methods are applied to a comprehensive dataset of granted patents and scientific publications spanning 2000 to 2023, sourced from Orbis intellectual property and Web of Science databases. Findings The analysis reveals a dual OI configuration: Baidu engages openly in scientific collaborations to foster value creation, while relying on centralized patenting strategies to secure value capture. This modular approach reflects a dynamic governance of knowledge across research and patenting domains. Baidu structures its AI innovation through selective openness, enabling agile adaptation in a rapidly evolving technological landscape. Originality/value This study contributes to research on AI, OI, business model innovation and dynamic capabilities by illustrating how hybrid openness strategies function as organizational mechanisms for sensing, seizing and transforming. It offers interpretive insights into the design tensions of OI and provides a grounded perspective on how firms strategically navigate collaboration, protection and innovation in data-intensive contexts.