In 2019 AI became a political, human rights and trade issue. Where does this leave China’s AI superstars?


  • Technology
  • Thursday, 26 Dec 2019

It was the first time human rights was cited as a reason for inclusion on the US trade blacklist. — SCMP

When Beijing declared plans to become the world leader in artificial intelligence (AI) in 2017, it alarmed the US and the rest of the world, according to former US secretary of state John Kerry.

In a conference in May, Kerry said Chinese president Xi Jinping’s announcement was not the “wisest” move: “It would have probably been smart to go try to do it and not announce it, because the announcement was heard in Washington and elsewhere.”

His words foreboded a storm approaching Chinese AI firms. Reports days later indicated Washington was considering placing several Chinese surveillance companies on the US Entity List, effectively banning them from purchasing core components from American companies.

Trump administration officials confirmed the news in October, announcing that eight Chinese companies – including national AI champions SenseTime, Megvii and Yitu – were to be added to the Entity List, along with 20 police departments.

“These entities have been implicated in human rights violations and abuses in the implementation of China’s campaign of repression, mass arbitrary detention, and high-technology surveillance against Uygurs, Kazakhs, and other members of Muslim minority groups”, the department filing said.

It was the first time human rights was cited as a reason for inclusion on the US trade blacklist.

By linking some of China’s most promising AI firms to alleged human rights violations, some observers said the US opened a new front in its tech war against China, signalling a willingness to use non-traditional means to contain China’s tech ambitions.

The politicisation of AI threatens to further widen the divide between the world’s two biggest economies, which are jostling for dominance in advanced technologies – from 5G to quantum computing – against the backdrop of a protracted trade war that has seen the countries impose hundreds of billions of dollars worth of tariffs on each another’s goods.

US President Donald Trump meets with China's President Xi Jinping at the start of their bilateral meeting at the G20 leaders summit in Osaka, Japan in June 2019. Photo: Reuters

Although the countries made an apparent breakthrough in trade negotiations this month with a phase one deal that covered a wide range of issues, including forced technology transfer from foreign firms in China, the trade ban for companies placed on the Entity List went unaddressed.

It is not just human rights allegations that Chinese AI firms are up against; US officials and big tech leaders have increasingly pushed for the US to more actively challenge the rise of China’s dominance in the field.

In February, US president Donald Trump signed an executive order directing the US government to prioritise AI in its research and development spending, months after he reportedly received a memo from then-Defence Secretary James Mattis that suggested the creation of a national strategy for AI to keep pace with China.

In September, Trump’s head of technology policy Michael Kratsios called for “collective power” from both the US government and America’s private sector to keep the United States ahead of China in the tightening race for global artificial intelligence dominance.

And last month, a report by the National Security Commission on Artificial Intelligence (NSCAI) urged the US government to invest more in AI, raising concerns about the progress China has made and declaring that the latter is investing more in development in this field than the US. Google’s former CEO Eric Schmidt, who led the government-commissioned panel, warned that China was ahead in facial recognition and financial technology specifically.

The US is still the world’s largest AI market, with a 57% share compared to China’s 12% of the US$37.5bil (RM154bil) global AI industry, according to the 2019 China AI Development white paper by technology research firm International Data Corp (IDC) and Chinese tech media outlet Qbitai.com.

China is fast catching up, however, becoming the second largest AI market in the world this year with a whopping 65% growth year-on-year.

The momentum has been driven by the 2017 national plan on AI, according to Andy Chun, an associate professor at City University of Hong Kong and convenor of the AI Specialist Group at the Hong Kong Computer Society.

China aims to catch up with the US on AI technology and applications and grow the value of its core AI industry to 150bil yuan (RM88.4bil) by 2020.

The national plan and the industry’s desire to meet the 2020 milestones defined in it are behind the rapid growth of China’s AI industry, Chun said.

Some of the banned companies have in recent years ranked among the strongest performers in the Face Recognition Vendor Test (FRVT), an influential US government accuracy test of facial recognition technology which comprises a number of different assessments.

The results of the FRVT are regularly cited by firms as a measure of their credibility, and are referred to by businesses and policymakers when buying facial recognition technology.

Dozens of countries – including Singapore, Thailand, the Philippines, Pakistan, Zimbabwe and Ecuador – have bought the Chinese firms’ products for uses ranging from police body cameras to “Safe City” CCTV networks.

But as pressure mounts from the US, China’s biggest trade partner, a shadow has been cast over the prospects of China’s AI firms.

In June, a United States senator proposed to bar companies from countries, including China, it said consistently violate “internationally recognised human rights” from the FRVT. If the proposed bill is passed, Chinese companies, already dealing with potential issues from having their supplies from the US restricted, could face further challenges remaining competitive in other markets.

So far, however, the impact of these developments on Chinese AI firms seems to be muted: many of the companies put on the blacklist had prepared for the worst since talk of the move first circulated in May.

Yin Qi, chief executive of Beijing-based facial recognition specialist Megvii, said the company was “well-equipped for the fight”, in an internal memo issued in early October and obtained by the Post.

Hikvision, China’s largest surveillance camera maker, cited the stockpiling of key components as a major reason for the 70% surge in inventory in the first nine months of the year.

Many Chinese companies have been looking towards developing more components and parts domestically to cut their reliance on foreign supplies.

Within a week of landing on the list, iFlyTek announced that two of its products, a teleconferencing and transcription service, ran smoothly on a China-made computer processing unit. Other companies such as SenseTime and Yitu have also been developing their own AI chips, holding out hope that over time China can develop home-grown equivalents to those produced by the likes of Intel, Nvidia and Qualcomm.

This reflects a move this year towards vertical integration, or owning all products and services in their supply chain, rather than their previous focus on horizontal expansion, or growing existing methods of business, according to Benson Ng, Ernst & Young Greater China digital advisory leader.

While undoubtedly a costly endeavour, this might notw be a bad thing in the long run, according to Ng.

“The US entity list will likely slow down the development of the global AI industry as a whole, but it will drive more opportunities in the development of AI algorithms and foundational technology in China,” said Ng, who added that he foresaw greater collaboration opportunities between China and markets in Europe, the Middle East and Africa in coming years.

Currently, China’s reliance on US-originated frameworks – such as Google’s TensorFlow and Facebook’s Pytorch – constitutes a significant gap in its AI ecosystem, which comprises foundational technologies like algorithms and frameworks, as well as data, semiconductors and computing power.

According to Ng, AI technology can be broadly divided into three categories – applications, algorithms and foundational components such as chips.

“We expect to see more vertical development of Chinese-made technology across applications, algorithms and foundational components in 2020 and beyond given trade tensions,” Ng said.

Going into 2020, the industry may also see more challenges in valuing start-ups and raising funds, according to Kai-fu Lee, a Chinese venture capitalist and chief executive of Sinovation Ventures.

Recalling how in 2017, at the start of China’s AI craze, an AI company’s valuation could be defined by the number of PhD holders among employees or even the number of awards it received, Lee said at a forum at the end of November that more investors in the sector were “returning to their senses”.

China’s AI industry is likely to see its first-ever downturn in both deal numbers and total venture capital and private equity investment in China AI since 2012 this year, according to real-time data compiled by Chinese tech research firm IT Juzi. The data showed that investment in China’s AI start-ups shrank from 124.3 billion yuan in 2018 to 84.1 billion yuan in 2019.

“Now the tide has gone out, you learn who’s been swimming naked,” said Lee, referencing a famous Warren Buffett quote.South China Morning Post

Article type: metered
User Type: anonymous web
User Status:
Campaign ID: 1
Cxense type: free
User access status: 3

Artificial Intelligence

   

Did you find this article insightful?

Yes
No

Across the site