Google, Apple spur Korea's voice AI push

South Korean telecom and internet companies are expanding Korean-language voice AI and multimodal services.

The move comes as Google and Apple push to transform voice assistants into AI agents capable of handling multistep tasks across apps and devices.

The speech recognition market is projected to rise from US$23.7 billion in 2026 to about US$104 billion by 2034, according to Fortune Business Insights.

SK Telecom has put its in-car agent A. auto into Renault Korea’s Filante while KT pushes its Genie TV AI agent and LG Uplus prepares a Malaysian launch for its call service ixi-O with Maxis.

Kakao has beta-launched its Kanana-O model and added voice actions in KakaoTalk, while Naver is testing its AI Tab search feature and upgrading Clova Note with speaker identification.

🔗 Source: The Korea Times

🧠 Food for thought

Implications, context, and why it matters.

Flat mobile revenue pushes telecom companies toward AI

South Korean telecom companies are moving into AI because their core mobile business is under strain 1.
Even with double-digit gains in 5G subscribers, mobile service revenue barely rose. KT posted 2.7% growth and LG Uplus 1.7% 1.
Average revenue per user (ARPU) is a telecom industry measure of how much money each customer generates. It is flat or falling. SK Telecom’s dropped 2.1% year on year 1.
That pressure is pushing carriers into new business lines. SK Telecom’s enterprise revenue rose 11%, helped by Internet of Things (IoT) services and cloud computing 1.
Industry chiefs are also asking the government for tax breaks and rule changes to speed up AI investment 2.

South Korea’s AI push offers a way to challenge US tech giants

Local companies are building “sovereign AI” models, which are AI systems built and trained at home for Korean language and culture 3.
The goal is better performance on local nuance. Naver’s HyperClova X used 6,500 times more Korean data than GPT-4, and SK Telecom said its A.X 4.0 model scored higher than OpenAI’s GPT-4o on local benchmarks 3.
This approach may help other countries build domestic AI leaders by focusing on local contexts that global models can miss 3.
It could also split the global AI market further, giving regional companies room to grow and pushing US tech giants to work with local partners.