China’s frenzy over ChatGPT-style services on full display at AI show despite Beijing’s caution
- Many exhibitors showcased LLM-powered applications in a wide range of sectors spanning logistics, education, finance, and construction
- SenseTime said on Friday that the company has completed a ‘comprehensive upgrade’ of its large AI model called SenseNova
China’s ChatGPT-frenzy was on full display at the country’s largest artificial intelligence (AI) fair in Shanghai this week, with local firms eager to show their services despite US sanctions on chips and Beijing’s cautious attitude towards generative AI.
At the World Artificial Intelligence Conference (WAIC), a three-day event organised by Chinese authorities to showcase the country’s AI ambitions and capabilities, chat bots and large language models (LLMs) were all the rage. However, ChatGPT, the flagship product of OpenAI, and Google’s Bard, were noticeably absent as China’s market is currently closed to foreign generative AI applications.
With OpenAI announcing GPT-4’s general availability to developers through the GPT-4 application programming interface (API), the WAIC floor was packed with Chinese LLMs and localised applications. On Friday, Alibaba Group Holding’s cloud unit unveiled Tongyi Wanxiang, a Midjourney-style application for generating images.
“The model supports multiple languages that allows users to describe what they want to present in the photo,” Zhou Jingren, chief technology officer of Alibaba Cloud, told the conference.
Zhou added that the model can generate additional photos in the same style of the original, if requested.
Alibaba’s Tongyi Qianwen, which generates text answers like ChatGPT, was demonstrated to visitors at the company’s booth. The bot offers specific functions, such as writing a love letter, teaching you how to praise someone, and cooking recipes. Alibaba owns the South China Morning Post.
Many exhibitors showcased LLM-powered applications in a wide range of sectors spanning logistics, education, finance, and construction. A JD.com executive told the conference on Friday that the Beijing-based e-commerce giant has invested over 100 billion yuan (US$13.8 billion) on AI, and has filed for over 20,000 AI-relevant patents.
JD.com is using LLM technology to help summarise and generate product listing details, a process that would traditionally require a lot of labour, the executive said.
