New challenge seeks LLMs for telecom network fault diagnosis, focusing on edge deployment and explainability.
The AI Telco Troubleshooting Challenge, launched November 25, 2025, by GSMA along with ETSI, IEEE GenAINet, ITU and TM Forum, calls on telecom operators, AI researchers, and startups to develop advanced large language models that can perform root cause analysis (RCA) of telecom network faults. Network outages are expensive and disruptive to telecom providers, and AI-driven root cause analysis aims to reduce downtime and improve efficiency. The competition evaluates models for their ability to generalize to new faults, deploy lightweight edge models, and provide clear explainability of their reasoning. The initiative builds upon curated datasets and benchmarking frameworks, including GSMA’s Open-Telco LLM Benchmarks. AT&T reported success with a fine-tuned 4-billion-parameter model outperforming leading AI systems, highlighting the potential of specialized LLMs in telecom operations.
Source: ETSI Newsroom