The release of ChatGPT late last year has driven global enthusiasm for artificial intelligence (AI) to new levels of frenzy. Overnight, the AI chatbot sparked a zeal among both industry insiders and the general public for equipping themselves with AI knowhow, lest they be left behind by the new revolutionary trend.

Driven by the AI-mania, many cast their eyes on this year's World Artificial Intelligence Conference (WAIC), which took place in Shanghai from July 6 to 8. First inaugurated in 2018, the annual WAIC gathers industry leaders, officials and experts to discuss the latest AI trends and developments. Over the past five editions of the annual conference, agreements have been signed for 200 major projects, with a combined investment of 70 billion yuan ($9.7 billion).

More than 400 companies participated in this year's event, a new record and double last year's number. They occupied a sprawling exhibition area of 50,000 square meters, more than three times that of last year. Chinese tech giants and startups rushing to join the global AI race are desperate to demonstrate their latest developments, discover new trends and explore business opportunities through the conference.

About 1,400 high-profile attendees shared their industry insights, among whom were world-renowned entrepreneurs, government officials, delegates from international organizations and top scientists such as recipients of the Association for Computing Machinery A.M. Turing Award, often referred to as the "Nobel Prize of Computing."

The atmosphere of AI permeated the entire city during the three-day event, which attracted more than 600,000 offline visitors. Parents and teachers took primary school students to visit the conference's exhibition, and some families even walked around the venue pushing strollers. "They are starting their AI-learning from babyhood," an exhibitor quipped.

Large language models (LLMs), robots, chips and autonomous vehicles were the mainstays of the conference, while dazzling new products offering theme-park-like interactive experiences kept visitors' senses in overdrive. Curious explorers had access to almost all things AI-related and insights from the guest speakers brought them up to date with the latest developments in the industry. All these made this year's WAIC the most visitor-friendly edition since its inauguration in 2018.

AI for industries

LLMs like ChatGPT, deep learning algorithms that understand and generate text in human-like fashion, dominated this year's WAIC. Exhibitors were largely either engaged in the research of LLMs or their application in other industries, Hu Houkun, Rotating Chair of Chinese telecommunications giant Huawei, said after touring the exhibition.

More than 30 LLMs took the spotlight during the event, including Huawei's Pangu, Chinese AI leader Baidu's Ernie, and Chinese leading AI software company SenseTime's SenseNova. According to a report released in May by a research center affiliated with the Ministry of Science and Technology, China has developed 79 major AI LLMs. While tech giants including Alibaba, Tencent, Meituan and iFLYTEK have all ventured into this field, small and medium-sized developers such as Langboat, Mobvoi and Transwarp have also followed suit, focusing on LLMs for specific industries.

Wu Hequan, a member of the Chinese Academy of Engineering, said at the WAIC that aside from applications such as conversation, poetry writing and painting, the LLMs should also be applied in real-economy fields including urban development, fintech, biomedicine and manufacturing.

Hu echoed Wu's opinion, saying that as the potential of AI is clear, it's essential for the way forward to be planned. Huawei's plan is to promote the in-depth development of AI and use it to empower industrial upgrading.

Huawei will help boost computing power to support the development of China's AI industry and make new innovations to allow AI to serve a wide range of industries as well as scientific research, according to Hu.

Huawei launched the latest version of its Pangu pre-trained deep learning AI model, Pangu 3.0, during the WAIC on July 7. Pangu can currently serve more than 10 industries including finance, manufacturing, government affairs, meteorology, electric power and healthcare.

Taking meteorological applications as an example, the Pangu-Weather AI model has demonstrated its higher precision compared with traditional numerical prediction methods for forecasts of one hour to seven days, making its predictions 10,000 times faster than through traditional methods, according to Huawei. The model can accurately predict temperature, humidity, pressure and wind speed and other conditions in seconds.

Wang Haifeng, Baidu's chief technology officer and Director of the National Engineering Research Center of Deep Learning Technology and Application, said at the conference that deep learning and LLMs have constituted a solid foundation of industrial intelligence, which will bolster the intelligent reconstruction of society on the whole.

"AI large models will eventually bring more innovation and possibilities to all industries," He Xiaodong, head of JD Explore Academy, said. The academy is a research institute of e-commerce leader JD.com that focuses on AI, quantum computation, data science and decentralized computing. He said in the future, people and machines will be more seamlessly connected and machines will be better equipped to help users accomplish greater tasks.

Tang Daosheng, Senior Executive Vice President of Tencent and President of Tencent Cloud and Smart Industries Group, underlined the importance of prioritizing the development of customized AI models. "A general LLM can solve problems on 70 to 80 out of 100 occasions, but it may not be able to 100 percent satisfy customers' needs on each occasion," he said at the conference.

To date, AI large models in China have shown great potential for application in fields such as smart mining, finance, intelligent manufacturing, railroad management, drug development, meteorology and e-government.

Greater input

According to a global innovation index report released at the WAIC, China ranks second in terms of AI innovation with progress in education, patent output and talent. "China's AI industry is in the midst of a development boom, with the output value of its core segments hitting 500 billion yuan ($69.4 billion) and the number of enterprises in the sector exceeding 4,300," Xu Xiaolan, Vice Minister of Industry and Information Technology, said at the conference's opening ceremony.

Nevertheless, industry insiders said the country, despite having a vast market for AI applications, still lags behind advanced countries in terms of basic technologies. Other challenges described in the abovementioned report include insufficient infrastructure and an underdeveloped AI innovation funding market.

"China has been accelerating the construction of AI-related infrastructure," Xu said, noting that its total computing power ranks second globally, while the number of its 5G base stations has exceeded 2.8 million. At present, China's 5G network is the most advanced and largest in the world.

She said, so far, more than 2,500 digital and intelligent workshops and factories have been created across the country, which has shortened the research and development cycle by 20.7 percent and improved production efficiency by 34.8 percent.

In a next step, China will focus on improving its computing power, algorithm and data processing capacities, consolidating basic technologies, optimizing its AI hardware and software and upgrading its AI industrial chain, which will in turn promote the technological and industrial development of AI, Xu said.

"Given the number of people working in AI-related fields, the advantage of a huge amount of AI data, the advanced AI infrastructure and the supportive policies rolled out by the government, China will become a world leader in AI by 2030," Liu Hongbin, Executive Deputy Director of the Center for Artificial Intelligence and Robotics at the Hong Kong Institute of Science and Innovation under the Chinese Academy of Sciences, said. BR

(Print Edition: Riding the Trends)

Copyedited by G.P. Wilson

Comments to zhangshsh@cicgamericas.com