ABSTRACT
Collaborating with one competitor is difficult but collaborating with several competitors is a monumental challenge. However, multi-competitor coopetition, or cooperation between multiple competitors, is increasing. This study examines how recent advancements in artificial intelligence (AI) and blockchain can support multi-competitor coopetition by enhancing governance. Examining two coopetitive R&D consortia in pharmaceuticals and medical imaging, we find that a nascent form of AI called federated learning can address key coopetition concerns such proprietary and confidential data protection, knowledge leakage, data sovereignty and silos thereby maintaining organisational boundaries and autonomy. The use of federated learning and blockchain increases transparency and accountability, which reduces information asymmetries and power differential inequities. Together, these technologies decentralise governance and authority, reducing the tension between collective value creation and individual value appropriation inherent in coopetition, particularly those with multiple competitors. Finally, this study illustrates how emerging technologies challenge traditional assumptions about organisational boundaries, distributed innovation, and coopetition.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Multi-competitor coopetition is defined as collaborative relationship among three or more competitors. Not all R&D consortia, alliances, standards groups, and innovation networks are coopetitive since they may not include multiple, if any, competitors. Similarly, multi-competitor coopetition includes three or more competitors, but may not be organised as consortia, standards setting groups, and innovation networks. Open innovation ecosystems include communities of firms united to use shared technology to create value (Chesbrough Citation2003; Vasudeva, Leiponen, and Jones Citation2020; Olk and West Citation2020), but this requires neither direct cooperation nor competition.
2 The third potential case was not included since it is primarily a university research initiative.
3 Examples of archival data sources include The Lancet, Nature, IEEE, Drug Discovery Today, Neural Information Processing Systems Conference, Journal of Machine Learning Research, International Conference on Learning Representations, Annual ACM Symposium, European Patent Office, and United States Patent Office.