Besides using a customized GenAI chatbot to help athletes find their way around, the IOC also relies on AI best practices

To help about 11,000 athletes of the Paris Olympics — all with varying languages and cultural backgrounds — to navigate the event venue and comply with rules and guidelines, the International Olympic Committee (IOC) has adopted a generative AI (GenAI) chatbot tailored for the task.

Developed in collaboration with technology partner Intel, the “retrieval-augmented generation” (RAG)  chatbot is capable of handling inquiries and interactions with Olympics athletes, offering on-demand information during their stay at the Olympic Village in Paris.

What makes this deployment a useful case study? Deploying GenAI solutions poses challenges like cost, scale, accuracy, development requirements, privacy and security. Since the IOC deals with sensitive personal information, its developers have to ensure that its technology partners offer highly secure and flexible, integrated yet customizable RAG components and solutions.

The solution the IOC accepted offers several key advantages:

  • The multi-lingual chatbot, called Athlete365, comprises an “industry-driven, out-of-the-box, production-ready RAG solution built on the Open Platform for Enterprise AI (OPEA) foundation”
  • It is designed to be highly flexible and customizable, integrating components from a catalog of offerings by multiple OEM systems and industry partners
  • The OPEA-based microservice components are integrated into a scalable RAG solution designed to deploy specialized AI-optimized hardware that allow it to scale seamlessly with proven orchestration frameworks (e.g. Kubernetes), and provide standardized application programming interfaces with security and system telemetry
  • Development of the large language model is easy on OPEA with hardware and software all optimized for the GenAI turnkey solution to facilitate open source, standardized, modular and heterogeneous RAG pipelines, focusing on open model development and support for various compilers and toolchains
  • The development system is part of the Coalition for Secure AI, which is an open ecosystem of AI and security experts from industry-leading organizations to share best practices for secure AI deployment and collaborate on AI security research and product development

By leveraging the advantages listed above, the IOC is making sure it is utilizing proprietary data securely, and enhancing the timeliness and reliability of AI outputs for landmark events such as the Olympics.