From left: Panellists Andy Chun, Scott Likens, Christopher Keil and Yam Ki Chan with moderator Barbara Meynert
Experts convene in Hong Kong for a lively discussion on the future of artificial intelligence
Will artificial intelligence save the world—or bring on the end of humanity? During the recent Asia-Pacific Business Forum (APBF) at tech incubator Cyberport in Hong Kong, four experts explored those questions and others.
At the APBF, presented by public agency Hong Kong Cyberport Management Co., the Hong Kong government’s Innovation and Technology Bureau and the United Nations’ Economic and Social Commission for Asia and the Pacific (ESCAP), many of the sessions focused on how the private sector can help reach the UN’s 2030 Sustainable Development Goals (SDGs).
Moderating the AI panel discussion, called The Artificial Intelligence Revolution: Implications for Work and Society, was Barbara Meynert, senior adviser with Hong Kong conglomerate Fung Group and chair of the ESCAP Sustainable Business Network’s Task Force on Digital Economy. To get things started, Meynert pointed out that in January at the World Economic Forum in Davos, Switzerland, Google chief executive Sundar Pichai said that AI is probably the most important thing humanity has ever worked on, more profound than fire or electricity.
A force for good
Yam Ki Chan, Google’s head of public policy and government affairs for Hong Kong and South Asia frontier, contended that the benefits of AI should accrue to everyone: “It’s going to help us solve messy problems that are critical to ending poverty and [meeting] the UN Sustainable [Development] Goals.”
First, Chan cited health care, which is using AI to help diagnose diabetic retinopathy, a disease that affects 450 million people worldwide. Second, he pointed to environmental protection. In Los Angeles, Chan said, a group of students are adding solar panels to old cellphones that will be installed in the forests of Peru and Indonesia to recognize the sound of chainsaws used for illegal logging.
Third, Chan highlighted AI’s role in accessibility. For example, YouTube is using machine learning to caption a billion hours of videos in 10 languages, “which will make educational videos accessible to the 300-million-plus people who are deaf or hard of hearing,” he said.
“Through this technology, we’re going to find ways to improve the conditions of certain groups that weren’t possible before,” Chan added.
Scratching the surface
Christopher Keil, managing partner with Silicon Valley investment and advisory firm TekStart LLC, said he looks at things a bit differently. “Probably the best line I’ve heard is that AI is not a part of the future of technology, AI is the future of technology.”
AI is now having such a wide impact that it’s changing the world, Keil said. It encompasses much more than artificial intelligence, he explained. AI is the outer layer, but the next layers down are deep learning, machine learning and neural networks. “We’re at a point where we are grabbing as much data as we fundamentally can. Then the deep learning starts, then the machine learning starts, and then we start training neural networks. We’re only touching the surface right now.”
Over the two years, only about 300 AI companies got venture funding, versus the 20,000 businesses that receive such capital annually, Keil noted. But Microsoft Corp. has pledged US$5 billion to AI over the next four years, and Chinese facial and image recognition specialist SenseTime Group Ltd. just raised US$600 million, making it the world’s most valuable AI startup. “I think we’re just on the cusp right now of the investment and the opportunity and what can actually be done with AI.”
“We’ve seen the maturity of AI in the last few years,” said Scott Likens, emerging technologies lead for the U.S., China and Japan with PricewaterhouseCoopers. Likens, who worked on the first web browser in 1994, recalled seeing the Internet come alive. “I wrote AI code 20-plus years ago,” he said. “It didn’t scare me until two years ago—and I say scare me in the sense of how fast it’s been moving.”
Likens has spent the past five years in China, arriving as the country was “building these amazing tech companies, which inherently created a platform of data for AI to thrive upon.” China’s current five-year plan is focused on building out AI, he said. “There’s a huge amount of money, but [also] a huge passion around this investment, and it’s moving so quick.”
In Asia, hundreds of thousands of people are doing jobs that AI will perform soon, Likens said. “I know 2030 is the date we have on some of these targets, but I don’t think it’s that far off, and I worry that we as a species have to adapt to it. The Internet was amazing, and it created this whole new economy over 10 or 20 years. AI is now creating this in two or three or five years.”
Jack Ma, founder of Chinese e-commerce giant Alibaba Group Holding Ltd., has said that in 30 years, the best CEO will be a robot, Likens noted. “The other lens is, should we worry about [AI], beyond just the workforce?” Likens said he sides with Tesla Inc. founder Elon Musk, who has been public about his fears.
PwC has stood behind having a responsible AI framework, he added, “working with governments to say, ‘How do we get ahead of this now and put up what I call guardrails?’ We want the optimization, we want the automation, we want the good side of what AI brings. But the speed of the advancement worries me and many others if we don’t put those guardrails in place.”
Likens doesn’t think the Skynet of the Terminator movies is coming anytime soon, but the combination of robots and so-called superintelligence gives him pause. He called for the development of AI that is free of human biases: “We can create unbiased artificial intelligence if we do it right and we do it now.”
Andy Chun, an associate professor of computer science at City University of Hong Kong, said his colleagues often ask him if AI will replace academics’ jobs in the near future. “Believe it or not, some people think it might happen very soon.”
By providing quality education for everyone, such a development could help reach SDG No. 4, Chun said. One roadblock to reaching that goal—today, about 263 million children and youth worldwide don’t attend school—is that teachers are in short supply, he noted. To achieve the UN goal by 2030, countries must recruit almost 70 million more teachers. “We need a seismic-level change,” Chun said. “Maybe technology plus AI can be that seismic change.”
Massive online open courses (MOOCs) are one cost-effective potential solution, Chun explained. “But what we’re talking about is quality education, and that’s where AI comes in.” Using technologies like big data, AI could analyze a student’s individual needs, personalize their learning experience and adapt that experience based on their needs and interests, Chun said. According to one scenario, by 2030, most teaching will be done by AI, he continued. “There will still be humans, but the humans are there just to manage the process,” Chun said. “I’m not sure if that’s going to happen, but there is a chance.”
PwC’s Likens said he has no doubt that AI will make jobs go away, but he predicted that new ones will emerge, giving data scientists as an example. By 2030, AI will create an additional US$15 trillion in global gross domestic product, his firm estimates. Half will be around automation, PwC reckons, while the rest will be around consumption—the ability for people to consume more. “Half of that $15 trillion—$7 trillion—is in China alone,” Likens said, observing that the country has aggressively moved from manufacturing into services.
But some people could be left behind, he acknowledged. “That’s where we really have to focus,” Likens stressed. “In emerging markets, markets that are very manual labour–based, you have to get ahead of that.”
Moderator Meynert asked if AI poses an existential risk to humanity, as Elon Musk has warned. “I’m not going to get political, as an American who lives in China right now,” Likens joked. “Because they’re listening.” People talk about the AI race being a bigger threat than nuclear war now, he said, and we’ve given up privacy for convenience. With AI now built into autonomous driving, “If I’m in an automated car that Google controls, do I trust them to move me freely?” Likens asked. “There’s an existential threat to us as people.”
For Likens, it comes back to being intelligent about how we train AI. “Most of the AI is training on how we shop and how we fly and how we eat. That is not only data I want it to see.”