SK telecom unveils AI multi-engine strategy for customized services

“Titan,” the supercomputer for the AI service A.DOT. [Courtesy of SKT]

South Korea’s leading mobile carrier SK telecom Co. Wednesday unveiled its artificial intelligence (AI) multi-engine strategy to offer various customized AI services to enterprises and public institutions by utilizing its own or third-party models of hyper-scale AI.

SK telecom said that it plans to integrate A.DOT, Claude2, and Konan as part of its AI multi-engine strategy. The idea is to provide carrier-based services such as chat and contact centers by linking its in-house A.DOT, Claude 2, developed by Anthropic PBC with a strength in ethical answers and large text processing, and Konan, which is rich in Korean data.

Claude 2, in particular, is known to be capable of inputting 75,000 words of text, equivalent to the size of a novel book.

SK telecom has set a goal of launching the A.DOT service for enterprises starting this month and the integrated customized service linking all three models from the fourth quarter this year.

“Starting with OpenAI Inc., many companies are preparing their own services,” the company said. “We have prepared a lineup that includes competent third-party large language models (LLMs) in addition to our own developed model.”

The mobile carrier has made investments to secure stakes in Anthropic and Konan Technology Inc. each.

Additionally, the company has established SAPEON Korea Inc., a semiconductor company, to resolve the issues of overheating and high prices of conventional graphics processing units (GPUs).

It has also invested in Persona AI Inc., an AI-based contact center (AICC), to expand its ecosystem.

SK telecom says that it is currently engaged in training and optimizing models for its customers. The plan is to use A.DOT for Korean-based LLM and Anthropic for cloud-based LLM.

“We are preparing to join the LLM-based pilot project being promoted by the government, including the Ministry of the Interior and Safety and the Ministry of Science and ICT,” the company said. “We are also actively pursuing LLM service construction projects in the financial sector with individual clients.”

The company is also in talks with a large manufacturing affiliate of SK Group to apply AI technology to the manufacturing sector.

“By combining our big data analytics business, an area where we have remained strong, we expect to contribute significantly to improving productivity through LLM AI technology, such as reducing production time for workers in the manufacturing process,” it said.