There’s a global race underway to capitalize on artificial intelligence, a cutting-edge industry that’s no longer science fiction. Canada, looking to carve out its place, is up against world economic superpowers. Reporter Emma Graney visited China to explore that country’s growing AI expertise and what Canada needs to do to compete. Her reporting was supported with a fellowship from the Asia Pacific Foundation.
Hong Kong — The conference crowd in Hong Kong is restless. Sophia, an eerily humanoid robot, is supposed to be talking to us, blowing our minds with the robotic prowess of hometown Hong Kong development company Hanson.
Instead, the robot needs a reboot.
For those at the Hong Kong AI Summit with nagging concerns about the notion of global robot dominance, it’s a welcome demonstration that, perhaps, we will not have to bow to electronic overlords any time soon.
Jeanne Lim is on stage with Sophia as the robot takes to blinking and smiling at the audience, rather than talking. Lim sighs, but her poise is unbreakable, even if this is not what she was hoping for.
Lim, chief marketing officer with Hanson Robotics, describes Sophia as a kind of robot pal, or maybe a toddler (who is currently throwing a tech tantrum of sorts, still not talking as planned).
“She’s like a little baby,” Lim says as audience members queue to take selfies with the robot. “That’s why I say I babysit her.”
To Lim, Sophia personifies artificial intelligence; when we think AI we often picture data sets and programs, and it all seems very removed and conceptual, she muses.
“But we feel that if AI (technologies are) going to be part of our lives, we have to humanize them. We have to be comfortable with them so it’s not this foreign thing that we put in the server room, but we’re used to them, we’re familiar with them, therefore we can influence them to take on our values,” she says.
Hanson is on the cusp of commercializing its technologies and producing robots for service industries, but that’s just one sliver of the artificial intelligence sector. When it comes to AI, Sophia’s cousins are already out in the world. They’re the apps on your phone, Siri and voice recognition, your smart fridge, surgery assistance technologies, self-driving cars and instant chats online with your bank.
Three weeks after Sophia’s tech tantrum, a mechanical barista slings coffee at the World Robot Conference in Beijing. It’s no Sophia — the body is rudimentary, there’s no wink, no smile, and certainly no banter about how your day has been — but by all accounts, it brews a fine latte.
Beyond the buzzwords
The AI Summit and World Robot Conference are but two of the umpteen AI-themed events in China and Asia each year.
It’s part of a global trend. This week in Montreal, for example, Canada will host a G7 AI conference as part of the Neural Information Processing Systems conference. The one-day event will see global thinkers and academics wrestle with how to foster societal trust in — and the responsible adoption of — AI, and build a common vision of human-centric AI. About 150 participants selected by G7 partners are expected to participate.
Professor Kam-fai Wong at the Chinese University of Hong Kong does his best to keep up with the rapidly changing sector. It’s hard, he says, because ever since “artificial intelligence” became buzzwords, companies have been clamouring to jump aboard.
Wong meets me at a small café in a Hong Kong mall. He’s a jolly man with a ready smile and neat hair who often punctuates his observations with a chuckle — the kind of guy who knows half the coffee shop.
“Anything on the ground at the moment, although they all claim to be AI, they’re not. It comes in cycles. This is the time for AI,” he says, leaning back in his chair.
Yet there is a real sense of urgency around investment and education, he says, particularly in China. He points to university research centres, departments and programs all solely dedicated to AI, and the government’s recent white paper outlining its goals to claim global AI supremacy by 2030. In particular, the economic giant is pitting itself against the United States.
“At this moment, the Chinese AI industry is still lagging behind, but they’re catching up fast,” Wong said.
It all boils down to talent — and money.
The U.S. has come to rely on immigrants keen to build a new life with more opportunities, Wong said. For the past 20 or 30 years it has built its power on the backs of hand-picked talent from across the globe, particularly China and India.
Despite the career-building gold offered by tech — education giants like MIT and Stanford, he says the allure of attractive salaries and benefits in China — like custom-built housing in Shanghai and Beijing — has many expats returning home.
Asked if the Asian superpower will meet its goals, Wong replies with an emphatic nod.
“In terms of national strategy, it’s inevitable,” he said, pointing to the “phenomenal” explosion of infrastructure over the past 20 years as an example of how quickly China can get things done.
“If you reference that kind of development pace … there’s no reason you can’t believe it would happen.”
Wong’s words ring in my ears as I walk through the World Robot Conference in Beijing, an annual showcase of all things robot.
More than a half-million people attend the four-day conference spread over hangar-sized exhibit halls at the Etrong International Exhibition and Convention Center. They come to check out robotic fish, squat home-surveillance machines on wheels, and singing educational robots with cherubic animated smiles.
Families dominate the event. Kids poke at games on robot faces and bop along with contraptions that companies like AI giant Ubtech hope will become the hot new thing — you’ve not seen robot marketing until you’ve witnessed a small electronic gadget breakdance with a human.
The conference also features a raft of forums, drawing academics from across the globe to talk about robots in public security, health care, commerce and elder care. A series of talks on next-generation AI delve into instinctive learning, making robots self-aware, and the similarities and differences between machine and baby learning. There’s even a day dedicated to discussions about robotic ethics and law.
AI in education
Artificial intelligence is an incredibly broad sector, but experts point to three major areas of rapid growth and investment: security (to be discussed more in Day 2 of this series), health care and education.
Take the pet project of Professor Dit-Yan Yeung at the Hong Kong University of Science and Technology. He’s passionate about making education accessible to as many people as possible.
“I have been using AI to help education in the sense that even online, people can provide personalized education,” he said in his Clear Water Bay office.
“Ultimately, I hope to provide everyone with a private tutor based on AI. Doing that, trying to democratize education with AI — that’s my mission.”
He’s not the only one.
Inside the Shanghai headquarters of Squirrel AI Learning, call-centre employees phone parents whose details are scooped from WeChat and other databases. Down the hall, a tutor goes through English exercises with her unseen pupil, drawing attention to grammatical intricacies.
A leg-up over the competition is vital in the hyper-competitive Chinese education system, where even preschools interview parents to see how their children will fit in.
Enter Squirrel’s adaptive learning algorithms.
Dr. Cui Wei, chief scientist with Squirrel, dashes into a restaurant in Shanghai’s Bund district. Bowing deeply, he apologizes profusely for his slight tardiness, explaining a previous meeting with investors ran later than expected. Straight after lunch he’ll hustle to another meeting to talk with developers about expanding subject offerings, he says, deftly snatching a slice of roast duck with his chopsticks.
Cui’s career has seen him shift from the quantitative finance world to adaptive learning in education. His eyes widen as he explains how one day he thinks Squirrel AI will compete with global educational tutorial behemoth Kumon.
It’s no pipe dream.
Adaptive learning means Squirrel can first survey students and pinpoint exact weaknesses, giving a huge advantage over traditional tutorial methods. Rather than relying on the brick and mortar after-school centres so common in Chinese cities, Squirrel executes targeted lesson reviews online, saving time, teacher costs and overhead.
Squirrel founder Derek Li is no stranger to the world of education startups. Everything about him screams modern entrepreneur — he effortlessly pulls off fashionably-cropped trousers with no socks and his face breaks into a smile of genuine delight when he meets you.
His was one of the first companies to build adaptive learning into online tutorials.
Sitting at a board table in his company’s offices, Li scribbles a series of staggering numbers on a notepad: The Chinese population is expected to spend around 270 billion renminbi (C$51.5 billion) on online educational materials next year, projected to grow by 20 per cent each year. A good chunk of that business will likely benefit Squirrel — an attractive prospect when you’re already increasing profits by around 40 per cent by nixing the traditional tutor role.
Li doesn’t see traditional teacher jobs diminishing (“The world will always have teachers,” he says), but he believes AI in education is the way of the future.
He illustrates his point by telling the story of his driver’s daughter, who was struggling in school. Her grades rocketed after Li set her up with a free Squirrel account, and now she’s been accepted into an engineering school.
“It works. It helps students. They learn,” he said.
“I think in the next 10 years, AI will change education.”
This article originally appeared in Edmonton Journal