SK Telecom said on the 25th it will join the "Global AI Red Team Challenge," held during MWC26, the world's largest mobile communications expo next month, to verify the safety and reliability of its self-developed artificial intelligence (AI) model A.X K1 (adot X K1).
The "Global AI Red Team Challenge" will test large language models (LLMs) developed by leading domestic and international telecom and AI corporations, with about 100 evaluators probing for potential misuse such as discriminatory responses and information breaches.
The event is co-hosted by the GSM Association (GSMA) and Selectstar, a corporations specializing in AI data and reliability evaluation, and will be held on Mar. 3 local time at the Talent Arena inside the MWC venue in Barcelona, Spain.
Because the "Global AI Red Team Challenge" is a program in which external experts directly inspect the safety of AI-generated responses, it is regarded as an event for corporations with strong technical completeness and reliability. The verification results produced by the evaluators are scored by a judging panel, and the number of participating corporations, final scores, and rankings are not disclosed externally.
SK Telecom will participate in the challenge with its self-developed large language model (LLM) A.X K1 (adot X K1). A.X K1 is a model independently developed across the entire process from architecture to training, and it was recently selected in the first phase of the national "independent AI foundation model project," officially recognizing its technical prowess.
The A.X model series, including A.X K1, is an AI model that SK Telecom has advanced based on data and operational know-how accumulated across its telecom and platform businesses. It is currently provided to customers through AI services such as adot, and its application scope is gradually expanding to various work areas within SK Group.
By taking part in this event, SK Telecom plans to validate A.X K1's performance and safety on the global stage and actively reflect the results in advancing its AI services to further enhance the completeness of its AI model.
Kim Tae-yoon, who oversees foundation models at SK Telecom, said, "As AI spreads widely into customers' daily lives and across society, ensuring safety and reliability is becoming more important alongside technological competitiveness," adding, "Through participation in the 'Global AI Red Team Challenge,' we will use this as an opportunity to further strengthen trust in SK Telecom's AI technology."