BEIJING - Transcribing is a big part of Mr Justin Wu’s work as a public relations executive in Shanghai, and the smart recorder of Chinese tech giant iFlytek is his best friend.

The recorder’s speech recognition technology cuts the time spent on transcribing by at least 90 per cent, said Mr Wu, 28. “The device can detect local dialects and accents that even I have problems differentiating,” he added.

All he has to do is connect the portable recorder to his mobile phone and the transcript will appear as it picks up what a person is saying. Mr Wu needs only to go through the text to make sure the transcript is accurate.

iFlytek first released the recorder in 2016, and its latest version costs at least 329 yuan (S$62).

Technology has become entrenched in everyday life in China, with artificial intelligence (AI) powering everything from transport to shopping.

In a 2022 survey by global market research firm Ipsos, 78 per cent of Chinese respondents agreed with the statement that products and services using AI have more benefits than drawbacks, but only 35 per cent of American respondents agreed.

Many workplaces in China are fitted with facial recognition scanners for employees, and such scanners also enable them to pay for purchases at convenience stores.

Cities around the country are rolling out robotaxis, which are in varying stages of development.

In Beijing, residents in some areas can hail these self-driving vehicles with a few taps on their smartphones.

The service is fully available for those in central Wuhan city and south-western Chongqing city.

Tech giant Baidu and autonomous driving tech company Pony.ai are among a handful of Chinese firms that have been given licences to operate robotaxis in China. Prices rival those charged by traditional taxis, varying by city.

In Yizhuang in the south-east suburbs of Beijing, fares start at 18 yuan. 

E-commerce platforms such as Taobao, Pinduoduo and JD.com have customer-to-manufacturer systems that use AI-driven behavioural data to inform factory production of everything from television sets to clothing.

According to global data firm Statista, the AI industry in China has grown from US$9 billion (S$12.3 billion) in 2020 to US$14 billion in 2022. It is projected to hit US$29 billion in 2023, rising to US$105 billion in 2030. This compares with a projected US$87 billion in the United States in 2023.

In terms of AI investment, innovation and implementation, China is No. 2 after the US, according to Tortoise Media, a British news site.

In 2022, for example, China secured US$13.4 billion in private investment funding in AI against America’s US$47.4 billion.

Stanford University’s 2023 AI Index shows that China leads in AI journal publications, producing nearly 40 per cent of the world’s total, ahead of the European Union and Britain at 15 per cent and the US at 10 per cent.

Growth of AI in China

China was behind the curve in the early decades of the development of AI, which refers to technology that allows computers to imitate human intelligence, resulting in machines that can perform tasks such as understanding language, recognising patterns and making decisions.

When the study of AI was taking off in the US and elsewhere in the West from the 1950s, China was in the throes of political turmoil, first with Mao Zedong’s anti-rightist campaign in the 1950s that saw intellectuals and scientists banished to labour camps, and then the chaotic Cultural Revolution of 1966-1976 that again saw the persecution of intellectuals.

In the late 1970s, however, science and technology were emphasised as a means towards economic development, with paramount leader Deng Xiaoping declaring in 1978 that they were “a primary productive force” for the economy.

In 1982, the Chinese Association for Artificial Intelligence was founded, and the study of AI started to take off.

But the use of AI really grew from 2011 after the development of big data, cloud computing and the Internet of Things, among other information technology innovations crucial to AI implementation.

The year 2011 saw the advent of WeChat, the micro-messaging app from Chinese tech giant Tencent that grew to be what tech venture capitalist and AI expert Kaifu Lee described as the digital Swiss knife for modern life.

WeChat was used to send text and voice messages to friends, pay for groceries, book doctor’s appointments, file taxes, unlock shared bikes and so on.

From 2012, start-ups focused on computer vision – another field of AI – burgeoned as demand for surveillance cameras in the security market grew. China’s Sputnik moment for AI happened in 2017, wrote Dr Lee in his 2018 book AI Superpowers: China, Silicon Valley, And The New World Order.

That was when China’s Go wunderkind Ke Jie, then 19 and the world’s No. 1 player, lost all his three games to the computer program AlphaGo, which was owned by US tech giant Google, showing how advanced AI was in the US.

AlphaGo’s victories over Mr Ke, wrote Dr Lee, “were both a challenge and an inspiration” to the Chinese, just as the launch of the Russian satellite Sputnik in 1957 was to the Americans, who went on to send astronauts to the Moon.

He noted that the AI frenzy in China had in fact already begun the year before, when AlphaGo defeated legendary Korean player Lee Sedol in a five-game match that was watched by 280 million Chinese viewers. “It lit a fire under the Chinese technology community that has been burning ever since,” he wrote.

In July 2017, less than two months after Mr Ke’s defeat, the Chinese government announced a development strategy to make China the world’s leading AI power by 2030.

Its biggest strength is that the government, private sector and academia are working hand in hand to achieve the AI dream, said observers.

Apart from an AI-friendly policy environment, the technology is also helped by the sheer amount of data that a country of 1.4 billion people can generate, given that AI gets better and more accurate with data.

The hunger and nimbleness of its entrepreneurs, its talent pool of AI scientists and the people’s ready acceptance of the technology in everyday life are further advantages China has.

Rivalry with the US

China’s intensifying strategic rivalry with the US, however, is threatening to curtail its ambitions.

The US had a head start in AI research. America’s innovators are helped by its unique research culture in which researchers spend years working to improve a technology without a product in mind, say analysts.

China also lacks the kind of unstructured data that foundation models, which are general-purpose AI systems that underpin generative AI such as ChatGPT, are trained on.

Just 1.5 per cent of all websites are written in Chinese while 56 per cent are in English, benefiting American model-builders.

In addition, the most advanced chips used to train AI are designed in the US and built with tools from America’s allies.

China is further disadvantaged by how it is unable to attract as much foreign talent as the US can.

Still, China’s rapid advances in AI are causing concern, particularly in the US and elsewhere in the West, that its government could leverage the technology to suppress its people and project its power beyond its shores.

There is also worry that it could use AI to stir dissent and divide the West through misinformation and influence.

In 2020, the US, on then President Donald Trump’s orders, started to block China’s access to cutting-edge computer chips necessary to the technology.

“US chip restrictions are already a significant problem for AI companies in China and that’s probably going to get worse in the coming years,” said Mr Matt Sheehan, a fellow at the Carnegie Endowment for International Peace who specialises in China’s AI ecosystem.  

“Their current stockpiles of advanced chips will dwindle, and the US may continue ratcheting up the controls even further,” he added.

So far, only Taiwan Semiconductor Manufacturing Company and Samsung, South Korea’s electronics giant, have been able to mass produce 3-nanometre chips, with plans to make even smaller ones by 2027, using Dutch company ASML’s equipment.

The US has limited the advanced technology from being exported to mainland China, whose main chipmaker Semiconductor Manufacturing International Corporation lags generations behind the most advanced technology.

On Aug 9, US President Joe Biden issued an executive order aimed at limiting US firms’ investments in initiatives and projects that support Chinese development of sensitive technologies, such as AI, citing national security concerns.

The new rules, expected to kick in next year, specifically target semiconductors, microelectronics and quantum computing. 

However, Mr Bo Zhengyuan, a partner at Plenum business consultancy in Shanghai, who studies the tech war between US and China, said the economic impact of the executive order “will likely be exceedingly marginal, if not entirely non-existent”.

He said American investment in the stocks of Chinese companies had already been on the decline due to anticipation of the restrictions and the sluggish Chinese economy. The restrictions may in fact provide US investors with clarity on which Chinese companies to invest in, Mr Bo said. 

The order states that the US “intends to keep pathways open for US investors to invest in Chinese AI technologies that do not raise security concerns”, and this could help avoid blanket prohibitions on investment in Chinese AI.

Beijing has taken tit-for-tat measures, announcing in July that it will impose export restrictions on gallium and germanium products used to make computer chips and other technological components, giving as reason protection of its national security interests. These restrictions kicked in on Aug 1.

Mr Bo said global talent feel pushed to choose sides.

China Centre for International Economic Exchanges vice-chairman Zhu Min said during the World Economic Forum’s “Summer Davos” in 2023 that China’s strengths are in its large market and how it is one of the largest data-owning countries in the world.

• OCBC is the presenting sponsor for the Asia Future Summit 2023. The event is also supported by Guocoland and Kingsford Group.

Currently, the US accounts for half of the big AI models launched globally, and China accounts for 30 per cent, he said.

Mr Sheehan said when it comes to US-China competition, Chinese companies have a strong track record of absorbing new advances or business models from abroad, and then adapting them for China and “squeezing them for all the value they can generate”.

“In terms of research, most people would still agree that the US produces the large majority of major advances in the field, but China produces lots of very good, if not elite, new research papers,” he added. “Chinese researchers have a lead in many image recognition technologies, and the same might be true for audio.”

The Chinese government is putting in place regulations to strengthen the AI industry by ensuring its sound and safe development. On Aug 15, for example, guidelines for firms developing generative AI – a form of AI that can produce original content like text, video and images, such as ChatGPT – came into effect.

Under the new guidelines, platform providers have to carry out security reviews and register their services with the government.

Companies that are not based in China but offer generative AI tools to Chinese residents must also comply with the rules, though Chinese firms with products targeted at markets outside of China are exempted.

Mr Sheehan said it is still too early to say whether the AI regulations in China would ultimately help its companies or constrain them. But if the rules create a stable regulatory environment, they could help accelerate Chinese AI progress in the long run.

For the everyday Chinese user of AI like Mr Wu, the PR executive, the future is exciting. As Chinese technology continues to improve by leaps and bounds, he believes China will only get stronger with fair and strong competition. “I don’t really approve of the way the US has limited the ways China can grow, but I have faith that my country can outdo the US,” he said.