The Chinese Tech Firms Pushing Boundaries Of Artificial Intelligence
In China’s quest to shed its reputation as a land of copycats, the world’s second-biggest economy is pouring resources in to the hottest area in technology innovation: artificial intelligence.
With the goal of nurturing world-class companies that can compete with the likes of Google and IBM in building intelligent machines, the Chinese leadership singled out AI as a key area of development in a report released during the National People’s Congress in March. Soon after, the country’s biggest technology companies — Alibaba, Baidu and Tencent — announced plans for AI laboratories and projects worth billions of dollars.
Many analysts believe AI is one area China can excel. With its huge number of online consumers, the country already generates vast amounts of data that could be used to understand behavioral and consumption patterns. The market for AI-related services, from chatbots to early-stage disease diagnosis, is projected to grow to 9.1 billion yuan ($1.3 billion) by 2020, from 1.4 billion yuan ($203 million) in 2015, according to Beijing-based consultancy iResearch.
China has already made real progress. In 2015, the country surpassed the U.S. with the number of published journals related to deep learning, a branch of AI that specializes in teaching machines to learn by themselves, according to the US National Science and Technology Council.
And their companies are following. Here are three Chinese AI companies shaping up to be leaders in their respective fields.
Forget about Apple’s Siri, or Amazon’s Alexa. In China, it’s iFlytek’s voice assistants that are powering household devices and business software. The Anhui-based company, which specializes in voice and speech technologies, has won many international competitions in speech synthesis and in translation between Chinese and other languages. It also beats Apple in speech recognition in the Chinese language, according to the Chinese Academy of Sciences.
The achievements mean that machines may soon take over typing and real-time translation jobs in China. Back in the 1990s, such tools were far less accurate than manual typing, with dialects, slang words and background noise tripping up the software. But recent advances led iFlytek to develop a cloud-based AI system capable of translating Chinese to a dozen languages including English, German and Uyghur. It has a 97% accuracy rate, according to the company.
iFlytek also enjoys a close relationship with state-affiliated entities. It is working with a dozen local governments, police and state-run banks to develop AI-based management and customer service platforms, according to a stock exchange filing. It is a market that is near impossible for foreign firms to break into, as local governments prefer working with domestic companies out of security reasons, said Ken Xu, a partner at Shanghai-based venture capital Gobi Partners.
“iFlytek’s focus on Chinese-language speech technologies, together with government’s protectionist bend, means it has a lot of future opportunities,” he added.
Another area AI is poised to make a change in China is healthcare, as the country faces a shortage of doctors. According to the World Health Organization, China has 1.5 physicians for every 1,000 people, compared with figures of 2 in Mexico and 3.5 in Australia. The situation means that Chinese doctors are overworked, underpaid and have experienced a string of physical attacks from disgruntled patients.
AI could help alleviate Chinese doctors’ workload, as the technology can be used for analyzing medical images, building healthcare databases and developing preventative treatments. One such firm is iCarbonX, a Shenzhen-based firm that is valued at $1 billion following an investment from Chinese web giant Tencent last year.
Wang Jun, founder of iCarbonX (Photographer by Calvin Sit/Bloomberg)
iCarbonX’s business starts from building a digital avatar for each person. By collecting biomedical samples such as saliva, proteins and DNA, the company wants to create a virtual version of yourself that could be used for healthcare tests, predict future diseases and recommend wellness programs. It is also partnering with seven biotechnology companies worldwide to analyze data, the company’s founder Wang Jun said during a January conference.
Such AI-related healthcare services are still at very early stages, as the firms need to collect enough data and develop a standard for the nascent industry, Xu said.
“We will not see a breakthrough in the next two to three years,” he said. “But progress will definitely be made for the next five to 10 years.”
The Chinese search engine operator, dubbed the Google of China, has dramatically ramped up AI research. It spends billions of dollars in R&D every year, and has said that AI accounts for the majority of those efforts.
It is Baidu’s driverless car projects that could ultimately make a difference, despite a string of recent setbacks including the departure of its chief scientist Andrew Ng, a top mind in AI who helped to found the Google Brain project in 2011.
The potential market for fully autonomous vehicle is huge in China. According to the Boston Consulting Group, the global market for vehicles with self-driving functions could reach as much as $77 billion by 2035, and China is expected to account for 30% of total sales. Chinese lawmakers are pushing for a unified framework of related laws too, driven by the desire to shift to a high-end economy rather than one built on investment and low-end manufacturing.
China’s own complicated road conditions may give Baidu an edge over the likes of Google and Uber. Every year, more than 260,000 people die in road accidents, according to the World Health Organization. Therefore, training robots to navigate such a difficult environment may lead to better technology, said Wang Shengjin, a professor at China’s Tsinghua University.
“Traffic conditions are much better in foreign countries, so their robots don’t have to cope with all those erratic driving problems,” Wang said. “But in China, researchers have to face these issues from the very start.”