SK Telecom and Deutsche Telekom to Jointly Develop Telco-specific LLM

South Korean wireless telecommunications operator, SK Telecom (SKT), and the German telecommunications company, Deutsche Telekom, have announced the signing of a Letter of Intent (LOI) to jointly develop a telco-specific Large Language Model (LLM), enabling global telecommunications companies (telcos) to easily and quickly develop generative AI models.

• Make Telecom Talk My Trusted Source
• 
• 

Also Read: SK Telecom, Deutsche Telekom, e&, and Singtel Launch Global Telco AI Alliance

LOI for Telco-Specific AI

According to the official statement released yesterday, the LOI signing ceremony took place at the SK Seorin Building in Seoul, with key executives from both companies in attendance. This development represents the initial step following discussions held by the Global Telco AI Alliance, which was launched by SKT, Deutsche Telekom, E&, and Singtel in July 2023, and it lays the foundation for their entry into the global market.

Also Read: Globe Telecom Explores AI to Enhance Customer Experience Through Automation

Development Large Language Model

SKT and Deutsche Telekom have announced their plans to collaborate with AI companies like Anthropic (Claude 2) and Meta (Llama2) to co-develop a multilingual large language model (LLM) tailored to the needs of telcos, including German, English, Korean, and more. According to the joint statement, the first version of the telco-specific LLM is scheduled to be unveiled in the first quarter of 2024.

Also Read: Airtel Deploys AI Based Speech Analytics Solution to Improve Customer Experience

Telco-Specific LLM

This telco-specific LLM is said to have a higher understanding of telecommunication service-related areas and customers’ intentions compared to general LLMs, making it well-suited for customer services, such as AI contact centres. The goal is to support telcos globally, including those in Europe, Asia, and the Middle East, in developing flexible generative AI services like AI agents according to their specific environments, ultimately saving time and costs associated with developing large platforms.