Ant Group has launched its own large language model (LLM) – the technology used to train chatbots like ChatGPT – and a new Web3 brand targeting Hong Kong and overseas markets, as the Chinese fintech giant boosts its capabilities in generative artificial intelligence (AI) for the financial services industry.
The Hangzhou-based company on Friday unveiled its self-developed “financial LLM” at the “Inclusion Conference on the Bund” event in Shanghai, along with two applications powered by the group’s AI model.
“We have built computing power at a level of 10,000 GPUs [graphics processing units],” Ant vice-president Wang Xiaohang said at the conference. “On this basis, Ant’s entire financial business has quickly switched to the LLM paradigm.”
Ant’s financial LLM was used to upgrade Zhixiaobao, the firm’s consumer-facing “intelligent financial assistant” for wealth management and insurance, to help users in various scenarios including generating asset-management plans. It also helped build Zhixiaozhu, an application that can assist finance practitioners in areas including investment research, insurance and marketing.
The upgraded Zhixiaobao will be available to users when it receives regulatory approval, while Zhixiaozhu is undergoing additional closed tests by Ant and industry partners, according to the company.
The new AI model of Ant, an affiliate of South China Morning Post owner Alibaba Group Holding, has already outperformed “mainstream general-purpose LLMs” in terms of financial scenarios, according to Wang.
LLMs are deep-learning AI algorithms that can recognise, summarise, translate, predict and generate content using very large data sets.
Ant’s foray into the LLM arena underscores the increased competition among China’s Big Tech companies to help the country close the gap with the West in developing innovative ChatGPT-like services, as the government sets out to implement a national standard for LLMs in line with efforts to regulate AI.
Baidu, Huawei Technologies, Tencent Holdings and Alibaba have already rolled out their respective LLMs for adoption in various industries, despite the heavy costs incurred from building massive computing infrastructure needed to train and develop new AI applications.
“It’s like throwing a stone into a pool: you are seeing the first wave, such as replacing customer operations, marketing and sales,” Jing said.
In the same Shanghai conference, Ant also launched a new Web3 brand called ZAN, targeting Hong Kong and other overseas markets.
This new operation headed by Zhang Hui, the former chief technology officer of AntChain, will provide a full suite of blockchain application development services for both institutional and individual Web3 developers. Zhang said ZAN would target developers who need security and compliance products.
Web3 is the hypothetical next-generation version of the World Wide Web that is decentralised and distributed through the use of blockchain and similar technologies.
Ant, which has pivoted towards hard tech after it came under regulatory scrutiny in late 2020, has doubled down on Web3 technologies through AntChain, a unit that focuses on blockchain, Internet of Things, data analytics and intelligent risk management.
In April, AntChain unveiled a series of new Web3 initiatives, including an open-source project featuring the company’s cross-chain technology and a Web3 research project called AntChain OpenLab.
“Ant firmly believes that technical services are our foothold in the Web3 industry,” said Ant’s Zhang. “We will work with our partners to solve the core issues of Web3 industry development.”
While the mainland has banned cryptocurrencies and taken a cautious approach to Web3, Hong Kong has been actively promoting the concept to burnish its fintech credentials.
The city’s Financial Secretary, Paul Chan Mo-po, earlier this year earmarked HK$50 million in the 2023-24 budget to speed up the development of the Web3 ecosystem.
Hong Kong’s push into fintech offers lessons for the world in generative AI, while the premise to develop Web3 is to “ensure the stability of the financial system”, Chan said in July during the World Artificial Intelligence Conference in Shanghai.
In April, Hong Kong agreed to collaborate with the municipal government of Hangzhou – capital of eastern Zhejiang province and hometown of both Alibaba and Ant – in areas ranging from technology innovation to Web3. South China Morning Post