[ad_1]
As synthetic intelligence (AI) expertise turns into extra refined and calls for for ethics and reliability develop, KT is accelerating its help for AI transformation (AX) in company and public sectors by selling its “accountable AI” framework.
KT introduced on Jan. 21 that it’s repeatedly strengthening its accountable AI framework and dedicating efforts to constructing an AI setting that firms and society can make the most of with confidence.
Lately, AI is evolving past easy content material era instruments into agentic AI that understands aims, makes impartial judgments, and performs duties. As AI’s involvement in decision-making and execution processes expands, industries are experiencing rising expectations for technological innovation alongside rising calls for for making certain AI security and reliability.
In response to those environmental modifications, KT has established AI ethics not merely as a compliance dimension however as a core worth for leaping towards AICT (AI+ICT)-centered administration, consolidating company-wide frameworks.
KT established the devoted group Accountable AI Middle (RAIC) in 2024 and have become the primary home telecommunications firm to nominate a chief accountable officer, CRAIO. Moreover, it established its personal AI ethics ideas known as ASTRI, making use of 5 core ideas all through the complete AI planning, improvement, operation, and utilization course of: accountability, sustainability, transparency, reliability, and inclusiveness.
KT additionally publishes its efficiency and AI ethics strengthening actions via the yearly launched “KT Accountable AI Report.” The report revealed in December final 12 months detailed KT’s self-developed AI mannequin’s social affect and security analysis processes, and AI guardrails (SafetyGuard). It additionally launched sensible instances comparable to RAI training growth.
KT continues to function the Accountable AI Advisory Committee composed of exterior specialists and participates in AI-related institutional discussions via cooperation with the federal government and trade. It’s also working to unfold accountable AI tradition by working AI ethics training as obligatory coaching for all staff, together with subsidiaries and accomplice firms.
On the technical aspect, KT reviewed home and worldwide AI danger administration frameworks to ascertain its personal security requirements appropriate for the home setting. In October final 12 months, it additionally revealed the “Accountable AI Technical Report” containing a full-cycle administration framework for AI danger identification, analysis, and mitigation, presenting analysis frameworks and implementation methods relevant to precise service environments.
The “AI Guardrail” that blocks dangerous responses from AI fashions in real-time was launched on the worldwide AI platform Hugging Face in September final 12 months and recorded an F1 rating of 97 factors in “Kor Moral QA,” a Korean AI ethics response high quality analysis commonplace.
Because of these efforts, KT’s self-developed AI mannequin Mideum Okay 2.0 Base turned the primary amongst home AI fashions to acquire AI Reliability Certification 2.0 (CAT 2.0) from the Telecommunications Know-how Affiliation (TTA). The corporate defined that this certification is critical because it represents the primary case of making use of a sophisticated certification framework primarily based on worldwide requirements, objectively verifying accountability and reliability all through AI mannequin improvement and operation.
KT’s accountable AI framework can be being acknowledged on the worldwide stage. The GSM Affiliation (GSMA) evaluates RAI maturity in 4 phases, and KT achieved the very best grade “Superior” (company tradition internalization) within the 2025 analysis, turning into the primary home cell telecommunications firm to be launched as a world finest apply.
KT plans to help home firms and public clients in pursuing AI transformation with confidence via reliable AI and readily relevant AI, past performance-centered AI.
KT CRAIO Government Managing Director Bae Quickly-min stated, “We’ll proceed to contribute to the wholesome development of the home AI trade and securing consumer belief primarily based on dependable AI expertise.”
[ad_2]
