Ant Group, the fintech affiliate of Alibaba Group, has announced that it is developing its own artificial intelligence (AI) model to power its chatbot service. The multimodal large language model, internally dubbed Zhenyi, is set to be used to enhance Ant Group’s ChatGPT-style service. This move places Ant Group among the growing list of Chinese tech firms competing with their Silicon Valley peers in the race to dominate the AI space. The company has already ramped up its scientific research efforts, investing CNY 20.46 billion in research and development in 2022: double its annual expenditure on such efforts in 2019.
Reports of Ant Group’s LLM work emerged soon after Beijing Shengshu Technology (Shengshu Technology) announced it had raised nearly CNY 100 million in its angel round of funding, with Ant being the lead investor. Incubated by Ant, RealAI (an AI solution provider), and BV (a venture capital initiated by Baidu), Shengshu Technology is an AI start-up under the leadership of Tang Jiayu, former vice president at RealAI, as CEO.
Ant Group has also signed a cooperation agreement with Tsinghua University, with a focus on two key fields: trustworthy AI and safe large-scale general-purpose AI models. Notably, Ant seeks to create partnerships with universities and colleges to further its scientific research efforts.
LLM is the AI technology behind OpenAI’s popular ChatGPT chatbot. Ant Group’s announcement reveals its intention to leverage in-house research to enhance its chatbot service. Social media users have speculated that the name Zhenyi was inspired by Wang Zhenyi, a renowned female scientist from the Qing Dynasty, known for her noteworthy works on astronomy and mathematics.
Going forward, Ant Group’s AI model development will put pressure on Silicon Valley giants to keep up with China’s burgeoning AI space. With increased investment in research and development, partnerships with universities and colleges, and the development of innovative AI technologies, China’s tech stars show no signs of slowing down.
