Telecom Giants Launch Open AI Initiative
GSMA, with backing from AT&T, AMD, and TensorWave, has launched "Open Telco AI," a global initiative to speed up the creation of AI models specifically for the telecommunications industry. The project aims to foster open-source collaboration to more rapidly deploy AI across telco infrastructure.
The initiative addresses a critical gap: general-purpose AI models often fail when applied to the specialized needs of telecommunications. These "frontier" AIs struggle to interpret complex network data, understand technical standards, and automate operations with the required accuracy. This performance gap is significant, with one GSMA Intelligence report noting that only 16% of generative AI deployments in the telecom sector have been applied to network operations. The Open Telco AI project aims to bridge this by creating a collaborative hub for developing and evaluating AI models specifically for telco tasks. AT&T is making a key contribution by releasing a family of open-source telco models. These models were trained on publicly available data and are designed to be agnostic to specific hardware or cloud platforms. Compute power for training, fine-tuning, and evaluating these new AI models is being provided by AMD and its cloud partner TensorWave. This involves access to their GPU platforms and open toolchains, which are essential for the heavy computational work of AI development. To measure the effectiveness of these new models, the initiative has developed the "Telco Capability Index." This benchmark will assess model performance across a range of telecom-specific tasks, providing a clear standard for what "telco-grade" AI can and should do. The collaboration extends beyond the founding members, with contributing partners like Huawei, KDDI, Nvidia, Orange, SK Telecom, and Swisscom, alongside several academic institutions. The project also has the support of other major players, including China Mobile, Deutsche Telekom, Google Cloud, IBM, Telefónica, and Vodafone.
