China’s SenseTime unfurls latest AI challenger to ChatGPT

HONG KONG – SenseTime Group on Monday showed off a suite of new artificial intelligence (AI) services developed with the company’s access to vast troves of data and deep computing power, including the latest Chinese challenger to

AI phenomenon ChatGPT.

Chief executive officer Xu Li took the stage during a virtual presentation to demonstrate the large AI model SenseNova and a user-facing chatbot called SenseChat.

Mr Xu, with help from staff, showed how SenseChat could tell a story about a cat catching fish, with multiple rounds of questions and responses.

Then he demonstrated how the bot could help with writing computer code, taking in layman-level questions in English or Chinese and translating them into a workable product. “With SenseNova, we can provide a supermarket of AI big models” for clients, he said. “We welcome our partners to connect to the big model and upgrade it with us.”

Mr Xu said human programmers now do about 80 per cent of the work in AI development, but in future, AI can handle 80 per cent of the effort while humans take on 20 per cent of the work to direct and polish.

The AI model can also help double-check, translate and revise code, he added.

Mr Xu also unveiled an image processing tool called Miaohua, or “draw in seconds”. A fourth demo about a product called Ruying, or “like a shadow”, can model the moves of a human to animate a digital being in a video.

The Hong Kong-headquartered AI software company also unveiled a service for modelling structures that use AI rendering to generate 3D buildings.

SenseTime, best known as a leader in computer vision, is joining a global race to develop generative AI since OpenAI’s ChatGPT captured the popular imagination.

Microsoft pledged a US$10 billion (S$13.3 billion) investment towards the US start-up, while rivals from Google to Baidu unveiled AI services that can similarly create original content from poetry to art just with simple user prompts.

SenseTime, which in March telegraphed Monday’s event by disclosing progress in training text-to-image large generative models, is also backed by Alibaba Group.

Like every other major tech firm, the Chinese online commerce leader founded by Mr Jack Ma is

working on integrating generative AI across its various services,

and began inviting corporate cloud customers to test drive the service last week. 

But there are concerns over whether Chinese companies can secure reliable access to the high-end chips and technology needed to develop large-scale AI models over the longer term.

SenseTime itself

is operating under US sanctions

that inhibit its access to capital as well as crucial American components, and the Biden administration in 2022 imposed curbs on the sale of AI accelerator chips to Chinese customers – a critical component in the development of any large-scale generative model.

SenseTime, co-founded by Massachusetts Institute of Technology alumnus Tang Xiao’ou, was one of the most highly anticipated debuts of 2021. It surged as much as 23 per cent on its debut, making Mr Tang briefly one of the world’s wealthiest people.

The firm has soared about 25 per cent in the days since news of Monday’s event surfaced on social media, firing up expectations among investors glued to every major AI revelation. But it is still more than 10 per cent below its debut price.

At the time of its initial public offering, SenseTime claimed in its prospectus to be the largest AI software firm in Asia, with an 11 per cent overall market share.

Its technology is used in a range of areas, including helping the police in China, providing product placements in films, even creating an augmented reality scene in a mobile game by Tencent Holdings. But its revenue began declining sharply in 2022 as the Chinese economy wobbled. 

Now, investors are hoping advances in AI can rekindle the sort of growth that once excited markets. Bloomberg