The accounting profession is both inspired and intimidated by the possibilities of artificial intelligence (AI). One of the most popular events at the recent World Congress of Accountants in Sydney (see At the crossroads of technology and integrity ) was a presentation on AI, which lauded its potential but warned against its pitfalls.
Kai-Fu Lee is a Taiwanese-American computer scientist turned venture capitalist. In his recently released book, AI Superpowers: China, Silicon Valley, and the New World Order, he looks at the differing approaches to the technology among developers in the United States and China.
Lee’s technology background enables him to easily explain to the reader the differences between neural networks, deep learning and AI. All of which are important distinctions in an era where the mass media tend to lump together various strands of new technology.
Tracing the history of AI and his own interest in neural networks from his time at Columbia University in the 1980s, Lee acknowledges that the fundamental research that allowed it to develop took place mostly in the U.S., Canada and the United Kingdom.
It was the victory over a human Go champion by the AlphaGo machine – a product of the British AI start-up DeepMind, which had been acquired by Google in 2014 – that became China’s Sputnik moment, galvanizing authorities into fast-tracking AI development, just how the launch of the Soviet Union’s artificial satellite in 1957 prompted Washington to redouble their efforts and eventually, win the space race.
Years of insights from international research enabled Chinese entrepreneurs and venture-capital funds – including Lee’s – to invest in this area. That funding, says Lee, leveraged “the country’s most significant strength: scrappy entrepreneurs with sharp instincts for building robust businesses.”
Today, Lee is confident that it is Beijing that has the AI edge, thanks to its enhanced ability to gather data about its citizens. “China has already surpassed the U.S. in terms of sheer volume as the number one producer of data,” he notes.
But it’s not just quantity that counts. “That [Chinese] data is not just impressive in quantity,” he adds, “but thanks to China’s unique technology ecosystem – an alternate universe of products and functions not seen anywhere else – that data is tailor-made for building profitable AI companies.”
While Lee charts the technological race between the U.S. and China in great detail, there is a deeper vein to the AI story: how far should the lines blur between human and machine. AI, he writes, has the potential to either uplift humanity or cast it into the depths.
Lee notes that even among the technologically savvy, AI provokes deep emotions. Google’s Director of Engineering Ray Kurzweil envisions a radical future in which humans and machines have fully merged. Tesla’s Elon Musk has described advanced AI as “the biggest risk we face as a civilization,” comparing its creation to “summoning the demon” during an interview at Massachusetts Institute of Technology.
The most pressing concern about AI – and one which accountants share – is its ability to displace human employment. Low-skilled and junior workers in particular, are concerned that machines can do their work.
Accountants, Lee argues, will potentially be in the firing line, along with assembly line workers, warehouse operators, stock analysts, quality control inspectors, truckers, paralegals and radiologists.
But Lee doesn’t see a bright-line divide between people skills and machine abilities. “AI’s biases don’t fit the traditional one-dimensional metric of low-skill versus high-skill labour,” he writes. “Instead, AI creates a mixed bag of winners and losers depending on the particular content of job tasks performed.”
Nevertheless, Lee’s perspective as a technology investor makes him hopeful about the future. Of the intense competition between American and Chinese AI innovators, he says: “This is not a new Cold War.” He admits that his use of the word “superpowers” in his title suggests conflict and confrontation, but notes: “AI today has numerous potential military applications, but its true value lies not in destruction but in creation.”
Although AI has the potential to disrupt society, Lee has no time for advocates of a coming “singularity” – when machines are undifferentiated from people. AI, he is certain, will remain in the hands of its creators for better or worse. “Our present AI capabilities can’t create a super-intelligence that destroys our civilization,” he writes. “But my fear is that we humans may prove more than up to that task ourselves.”
Author interview: Kai-Fu Lee
The career path of Kai-Fu Lee runs almost parallel with the history of modern consumer technology. He ran the multimedia division at Apple during the company’s wilderness years in the 1990s, worked at SGI, and started Microsoft’s China-based research unit before launching Google China in 2005.
It is that dual perspective – working for Silicon Valley behemoths in a China context – that has given him a unique perspective on how the two technology cultures compete. “It’s almost like a parallel universe,” he says, speaking on a recent book tour.
“The United States is very well-segmented, with companies like Google, Amazon and Facebook each having clearly… separate [roles]. In China, everybody was competing with everyone. Alibaba had the entire payment [ecosystem to itself] but all of a sudden, in one year, Tencent took almost half of that away from them.”
Born in Taiwan, Lee emigrated to the U.S. as a child and attended New York’s prestigious Columbia University before obtaining a doctorate at Carnegie Mellon University in Pittsburgh. He quickly became caught up in China’s technological entrepreneurial boom and counts Alibaba founder Jack Ma as a long-term mentor.
One of Lee’s hopes in writing the book was that it would put to rest some of the myths about artificial intelligence. “AI is the study of how to make machines think like us,” he explains. “With large amounts of training data, it can make very accurate decisions, predictions, classifications and syntheses, but only in one domain – that’s its Achilles heel.”
That, Lee reassures, is why you don’t have to worry about being ruled by cyborgs. “We humans can think broadly, strategically, creatively and cross-domain. “We have common sense and AI doesn’t have any of that. It is basically a pattern recognizer.”
Despite the potential for labour and social problems, Lee is convinced that AI will eventually be seen as of benefit to society as a whole. “When we look back in 50 years, I think we would recognize that AI is actually serendipitous, because it takes away the routine jobs that were never meant for us to do.”
AI, he adds, is humanity’s chance for freedom. “After thousands of years of evolution we are still here like little rodents running around in the wheel doing the same thing over and over again thinking that’s the meaning of our lives,” says Lee. “I’m thinking that our maker perhaps got so sick and tired of it he decided to give AI to us.”
