AI expands territory as LLMs embed into Chrome, KakaoTalk platforms

(Google)

Large language models (LLMs) such as GPT and Gemini are rapidly spreading across a wide range of platforms, as tech companies pursue cross-platform integration strategies to stay competitive in the intensifying artificial intelligence race.

Big Tech firms are not only embedding LLMs into their own ecosystems but also partnering with external players to broaden adoption, a shift seen as a strategic move to enhance user retention and expand reach.

Gemini in Chrome, GPT heads to Apple

According to industry sources on Sunday, Google recently integrated Gemini directly into its web browser Google Chrome. Users can now input queries into the address bar and receive real-time summaries and analysis while browsing, as the AI recognizes webpage content and responds accordingly.

Google has already extended Gemini to its video platform YouTube, in a broader effort to reduce user churn by embedding AI across its services, industry sources said.

Meanwhile, Apple, often seen as lagging in AI development, is seeking a rebound through partnerships. The company announced plans to integrate both Gemini and GPT into its voice assistant Siri, marking a shift away from its traditionally closed ecosystem.

This pivot reflects a pragmatic approach, allowing Apple to save time and cost on in-house AI development while immediately enhancing device competitiveness with proven models.

For Google and OpenAI, the collaboration offers access to hundreds of millions of users through Apple’s hardware ecosystem, making it a mutually beneficial strategy.

(Kakao)

In South Korea, Kakao has also joined the convergence trend. While developing its own model, Kanana, the company has introduced services powered by GPT within KakaoTalk, adopting a hybrid approach that balances proprietary development with proven global technologies.

Cloud evolves into AI department store

Cloud platforms such as Microsoft’s Azure and Amazon’s AWS are positioning themselves as “AI department stores,” offering a range of models from providers including OpenAI, Anthropic’s Claude, Meta’s LLaMA and Mistral AI.

Companies are increasingly mixing and matching models based on cost and performance — using lower-cost open-source models for simple tasks such as summarization or translation, and premium models for complex data analysis and coding.

The growing fear of a so-called “SaaS apocalypse,” in which traditional subscription software risks becoming obsolete due to generative AI, is further accelerating cross-platform integration.

Analysts say the success of LLMs will ultimately depend on how widely they can expand across platforms, as broader integration translates into larger user bases.

“LLMs are becoming a general-purpose technology like electricity or the internet,” an industry official said. “Rather than sticking to a standalone approach, companies that create the most efficient user experiences through convergence will define future competitiveness.”