This session showcases state-of-the-art advances in live sports production through the integration of artificial intelligence, augmented reality, and private 5G infrastructure.

Our first paper discloses an AI-powered assistant used in the Paris 2024 Olympics, which delivered ultra-low-latency, real-time computer graphic overlays for swimming and fencing. The system successfully demonstrated advanced scene understanding (including leveraging automotive lane tracing) whilst maintaining human-in-the-loop editorial control.

Next, a private 5G deployment during the 2025 World University Winter Games is presented, illustrating how 5G Non-Public Networks (NPNs), in combination with LEO satellite systems, can simplify live TV production in challenging environments including an indoor arena and an alpine skiing venue. Key performance metrics such as uplink throughput and latency, video signal resilience and ease of set-up are analysed.

Finally, a breakthrough AI/AR system for live outdoor cycling is introduced, overcoming long-standing limitations in these challenging production environments. This impressive work combines cyclist detection and tracking, 3D pose estimation, and automated point-of-interest recognition to generate contextually aware augmented visuals, all within a fully automated pipeline.