Tech giants are seeing the biggest gains from GenAI by integrating LLMs into apps and services that already had large user bases, according to QuestMobile. — SCMP
Artificial intelligence (AI) adoption is primarily benefiting users in three areas – social and entertainment, education, and office collaboration – according to a new report, highlighting where the country’s technology giants may see the biggest gains as they race to integrate generative AI (GenAI) into nearly every part of their businesses.
In their eagerness to capitalise on the rapid development of large language models (LLMs) – the technology behind GenAI products such as ChatGPT – tech firms have significantly expanded the use of AI-driven features across a variety of service offerings, mobile internet market consultancy QuestMobile said in a report published on Tuesday.
Big Tech firms such as Internet search giant Baidu, WeChat operator Tencent Holdings, TikTok owner ByteDance and ecommerce behemoth Alibaba Group Holding, owner of the South China Morning Post, stand to gain the most by integrating AI into products that already have large user bases, according to the report.
Alibaba fintech affiliate Ant Group has one of the most popular AI applications with its “intelligent financial assistant” called Zhixiaobao, which offers help with wealth management and insurance. The assistant had nearly 60 million users at the end of June, according to QuestMobile.
Baidu has been baking its Ernie LLM into a number of services, including its search engine and online document sharing service Wenku. Similar to what Google has been doing with Gemini, Baidu search now gives AI-generated responses to user queries at the top of the results. Wenku can generate slide decks based on simple text prompts. These were the No 1 and No 3 most popular AI features in China, respectively, according to a ranking by aicpb.com, which tracks the popularity of AI services.
The versatility of LLMs has driven the current wave of AI adoption, according to the QuestMobile report. LLMs are trained on vast troves of data, which allows them to generate sophisticated results from queries in plain language. Combined with other GenAI technologies, they can produce text, images, audio and video responses, often with surprising results.
GenAI is also known to produce false information, or in the case of China, politically sensitive information. The country requires government approval for the public release of LLMs.
A total of 188 models have been approved for release so far, according to the lists published by the Cyberspace Administration of China (CAC). They have been used to power apps and services across 20 main areas, including emotional companionship, artistic creation, online shopping and financial services. Nearly 600 million people in China are using these LLMs, the head of the CAC recently said.
GenAI is also having an impact on hardware, the QuestMobile report notes, as smartphone and personal computer vendors turn to AI to lift sales.
AI-capable phones – typically those capable of running AI models on device by using processors dedicated to the task – are expected to make up 54% of the global smartphone market by 2028, while 80 per cent of PCs priced above US$800 will be AI-capable by that year, according to Canalys, a market consultancy. – South China Morning Post