Junko Yoshida sees the writing on the Shenzhen Airport wall
The most convincing indication of China's rise in the AI market that I recently came across was the plethora of AI-driven toys displayed, demonstrated, and pitched at shop after shop in the Shenzhen airport.
China’s big AI plan — to dominate the global market with its AI technology by 2030 — is widely known and deeply feared in the rest of the world. The lingering mystery, though, is how much of the best-laid plans of China will come to pass, if ever.
China’s State Council proposed in July 2017 “A Next-Generation Artificial Intelligence Development Plan.” It says that China will ultimately become the world leader in artificial intelligence, with a domestic AI industry worth almost $150 billion by 2030. The first step is to catch up with the United States on AI technology and applications by 2020.
It’s easy to call this mere bravado, but it’s probably not wise. As I see it, the Chinese genie is already out of the bottle.
Beijing’s “national champion” technology leaders — Baidu, Alibaba, Tencent, and iFlyTek — are established, successful firms. They are leading the development of innovation platforms such as self-driving cars, smart cities, computer vision for medical diagnosis, and voice intelligence. The newest “national champion” listed by Beijing is the Hong Kong startup SenseTime, which specializes in face- and image-recognition technology.
Veteran reporters on the global IC beat remember a series of “five-year plans” launched by the Chinese government a few decades ago, explicitly intended to kickstart an indigenous semiconductor industry. These early efforts resulted in the establishment of SMIC but not much else. Then in the 2010s, a host of Chinese chip startups sprang up, each designing its own SoCs with licensable cores from companies such as Arm and Ceva. To date, the Middle Kingdom, however, hasn’t exactly cracked the global chip market, although China is now making an even bigger investment fund available for the domestic chip industry.
China’s AI plan and its big fund for the IC industry share common threads that include the government’s huge budgets, aspirations, and ego. But the AI plan is likely to follow a path of its own because China already has established AI giants.
Furthermore, the South China Morning Post, owned by Alibaba, claims that China now dominates AI funding. Last year, 48% of total equity funding of AI startups globally came from China, compared to a U.S. outlay of 38% and 13% by the rest of the world. This is a significant jump from China’s 11.3% in 2016.
The most convincing indication of China’s rise in the AI market that I recently came across was the plethora of AI-driven toys displayed, demonstrated, and pitched at shop after shop in the Shenzhen airport. The toys range from battery-operated “smart” dogs and dancing programmable robots to conversational C3POs and interactive AI “educational” robots with a remote “babysitting” mode that takes snapshots of the kids and transmits them to Mom and Dad.
AI-driven toy robots (Photo: EE Times)
You may laugh. After all, we’re only talking about toys whose AI capabilities are primitive. I laughed, too, until it dawned on me that China’s booming toy business is the rare — possibly exclusive — example of making real money with AI.
China’s “educational” robots come with a level of intelligence equivalent to that of low-end smartphones.
But it’s presumptuous to judge these so-called AI-driven consumer goods as just toys. An interactive chatty robot with an AI system equipped with a range of English-language learning materials is a cheap shortcut to fluency for Chinese parents who want their kids to grow up and speak English.
Even a robot whose conversation might seem a little dumb to a grownup doesn’t bother kids as long as the robots respond to them in an animated, interactive fashion.
My stroll through the toy arcade in Shenzhen airport reminded me of my first interview with Toshitada Doi, an inventor of Sony’s original robotic dog, Aibo, in Japan in the late 1990s. I asked why Sony chose to develop its first robot as a toy rather than something more useful. Doi replied, “People are more forgiving of toys.”
More AI toy robots (Photo: EE Times)
Knowing what’s “good enough” is one of the keys to effective product development. A technology that’s good enough to entertain little kids will, in a few generations, be good enough to perform for adults.
In the 1980s, for another example, you might recall solar-powered calculators. These handheld calculators powered by solar cells, or photovoltaics, became the first successful application of “solar on the cheap.” Solar panel experts in those days regarded pocket calculators made in Japan as “toys.” But amorphous silicon — the basic material of those primitive early photovoltaic cells — eventually became an integral element in LCD development, by Sharp, a few decades later.
I’m not saying that AI-driven toys will lead to China’s world AI domination. But given that AI’s business growth depends on a myriad of applications, the toy segment is as logical a gateway as any other.
I haven’t even mentioned China’s huge AI-intensive surveillance market. While these applications are controversial, they have proven a fertile ground for China’s AI startups.
Applications decide AI's growth
In Shanghai last month, I bumped into Weijie Yun, managing partner of a venture capital firm called Tyche Partners, and asked him about his firm’s strategy for AI technology investment.
He asserted, “I am not so much interested in investing in AI chip startups. I am looking for AI applications.” Asked why, Yun said, “AI covers a vast market, very fragmented by so many different applications. One AI chip can’t cover them. It means you need an AI hardware tailored to a specific application, thus making AI chips very hard to scale.”
Asked for an example of his VC’s AI investment, Yun talked bout AEye, Inc. (Pleasanton, California). AEye makes a system called iDAR. According to the startup, “iDAR takes solid-state agile LiDAR, fuses it with a low-light HD camera, then integrates artificial intelligence to create a smart perception sensor that is fully software-extensible.”
The startup faults the current generation of LiDAR for lacking the intelligence to discriminate as it gathers information. By applying AI to the process of collecting, producing, and assembling a huge amount of data, AEye believes that it can streamline the processing and computational power of LiDAR.
iDAR is no toy. But it’s another field to which AI can be applied for better results.
In other words, success in the AI marketplace doesn’t mean yet another new AI chip or algorithm. Knowing where to apply AI and do it well in the real market could be the first step for China to build its domestic industry — regardless of whether China hits its goal of $150 billion by 2030.
— Junko Yoshida, Global Co-Editor-In-Chief, AspenCore Media, Chief International Correspondent, EE Times