A Kakao data center in Ansan, Gyeonggi Province / Courtesy of Kakao
Korea’s mobile platform giant Kakao said Thursday it has formed a strategic partnership with Google to collaborate on-device artificial intelligence (AI) services and embed its technologies into the latter’s new extended reality (XR) devices, including AI glasses.
Kakao CEO Chung Shin-a announced the partnership during the company’s earnings call, saying the company will “begin full-fledged global collaborations to introduce next-generation AI experiences on the device side from this year.”
“As the starting point of the partnership, we will collaborate with Google Android to upgrade on-device AI services, including Kanana in KakaoTalk launched last year,” Chung said.
“Since this is a cooperation with the Android team, we expect to explore various opportunities in on-device AI services that can maximize the value of data assets within the Kakao ecosystem.”
Kanana in KakaoTalk is a chatbot-type on-device AI service that functions in Kakao’s flagship KakaoTalk messenger app. Based on the company’s own AI model, it understands users’ conversation context and proactively suggests schedule briefings and information updates, as well as recommending products. Because it functions within the smartphone, it has strengths in protecting user privacy.
Kakao said it is working with Google to optimize the service so that it can run smoothly on Android mobile devices. Kanana in KakaoTalk has been in beta service since October last year and is scheduled for an official launch in the first quarter, including on Android.
Kakao CEO Chung Shin-a speaks during a media day event in central Seoul, Feb. 4, 2025. Korea Times photo by Choi Joo-yeon
Chung also unveiled a partnership on XR devices, including AI glasses. Kakao said it will focus on establishing interfaces for new form factors to enable user experiences such as hands-free messaging and natural language-based interactions.
Along with that, Chung said Kakao is in talks with Google Cloud for an operating cloud computing system based on the latter’s tensor processing units (TPUs), in a move to reduce costs stemming from running AI services on graphics processing unit-based infrastructure.
Source: Korea Times News
