Integrating Decentralised AI Services in a Collaboration Ecosystem

Abstract

The increasing variety of generative AI systems and tools, particularly large language models, poses significant challenges for integration into larger software infrastructures due to issues with interoperability, scalability, and data protection. To address these challenges, this paper proposes a decentralised platform ecosystem that enables the composition of AI services from distributed data sources and models. The platform prioritises flexibility, scalability, re-usability, cooperation, and data protection, providing standardised interfaces and best practices for seamless collaboration among multiple stakeholders. By facilitating the creation of open, modular, and federated architectures, this ecosystem aims to unlock the full potential of AI in collaborative environments while ensuring secure data exchange and protecting sensitive information.