SHANGHAI (Reuters) – Chinese technology companies, from industry giants to ambitious startups, gathered this week at the World AI Conference in Shanghai to showcase their latest innovations and voice strong support for the country’s artificial intelligence sector, even as it faces U.S. sanctions.
More than 150 AI-related products and solutions are on display at the conference, with a small number of foreign companies such as Tesla and Qualcomm joining the mostly Chinese lineup, according to the event organizer.
The conference also served as a launching pad for some companies to showcase their latest AI products.
A notable example is SenseTime, which previously focused on facial recognition technology but recently shifted its focus to generative AI following OpenAI’s release of ChatGPT in late 2022.
On Friday, the company unveiled its SenseNova 5.5, its most advanced large-scale language model (LLM), which is billed as a rival to OpenAI’s GPT-4o in areas such as mathematical reasoning.
Despite the challenges posed by U.S. sanctions that limit access to advanced chips, many executives at the conference expressed confidence that Chinese companies will continue to thrive in the AI sector.
Zhang Ping’an, head of Huawei’s cloud computing unit, told a forum that the idea that a shortage of the most advanced artificial intelligence chips would hamper China’s goal of becoming an AI leader should be “abandoned.”
“No one will deny that China’s computing power is limited,” Zhang said. “If we think that not having the most advanced AI chips means we can’t be the leaders in AI, then we have to abandon that view.”
To address the lack of access to advanced chips, Zhang called for more innovation in areas such as cloud computing.
Liu Qingfeng, chairman of artificial intelligence company Iflytek, which like Huawei has been placed on a sanctions list barring it from buying advanced chips from U.S. companies, echoed those comments.
On the WeChat account of state-run news agency China News Service, Liu said in an interview that many LLMs developed by Chinese companies, including Iflytek’s, have been able to compete with OpenAI’s GPT-4.
“We must have our own LLMs, independently developed and controlled by China, capable of comparing with the highest global standards and continuously narrowing the gap,” he said.
Some executives, including the CEO of Chinese search engine giant Baidu, Robin Liurged the AI industry to move away from developing LLMs, which require huge amounts of computing power and AI chips, and instead focus on embracing AI.
“Without applications, having only fundamental models, whether open source or closed source, is useless,” Li said at the conference.
(Reporting by Casey Hall in the Shanghai and Beijing newsroom; Editing by Anne Marie Roantree and Anil D’Silva)
