Last autumn, at an official TED conference in Vancouver, Artificial Intelligence (AI) expert, Kai-Fu Lee stood on a red-carpeted, circular stage, the lights in the room focused on his dignified frame, the crowd around him sober and attentive.
Bespectacled and donning a blue round collared suit, he kicked off by announcing that he wanted to talk about out how AI and humans can co-exist. “But first,” he quickly added, “we have to rethink our human values.”
Lee is not a philosopher or a moralist but a man of science who has been standing at the front line of AI for the past three decades, as a researcher, manager at world-class technology firms and venture capitalist. So it would not have been a surprise if some had wondered why he was starting a talk about AI with an exposition on human values.
The TED talk, among other things, was a preliminary preparation for the launch of Lee’s latest book, ‘AI Super powers: China, Silicon Valley, and the New World Order’, which was published months later in September. There, Lee expounded on the key ideas he had shared during that TED talk. AI, he believes, will dramatically change the way we engage with reality and the only way to ensure the flourishing of the human race is to focus on things that computers cannot comprehend, things that are intrinsically tied to human values.
Artificial Intelligence is the new cool: everyone is talking about it, and everyone wants to be involved. “If we want to be one of the leaders of the upcoming world, we have to work on artificial intelligence,” French president Emmanuel Macron told Fareed Zakaria earlier this month. Other national leaders, business executives have and will continue to make similar statements. Ambitious governments, like the United States and China, have even developed comprehensive policy plans to take advantage of the impending economic windfall offered by AI.
It was John McCarthy who first coined the term ‘AI’ in 1956, when he held the first academic conference on the subject. McCarthy was part of a generation of pioneers (Alan Turing, Marvin Minsky) starting from the 1950s who started to seriously engage with the idea of recreating human intelligence in a machine. It is fairly easy to underestimate the complexity of such an endeavor, especially if the brain is imagined as a rule-based, procedural organism. Turing, who is largely considered as the father of Computer Science, once suggested that recreating human intelligence could be achieved if computers had enough storage capacity. Today, compared to Turing’s days, we are in computer memory heaven, yet AI’s ultimate goal is still a mirage.
In his book, AI Superpowers, Lee writes about the two broad approaches taken by AI experts, the rule-based approach and the neural networks approach. For a long time, the latter was mostly ignored since it required large swathes of computing power and data to be effective, two things that were once very expensive and scarce to come by.
An easy, stripped-down way to think of the neural networks approach is to think of how human babies learn about the world: by feeling their way through it using their five senses, making mistakes, remembering those mistakes and implementing corrections accordingly. A collection of these corrections is what we describe as ‘common-sense’. In the same way, data about the world is fed into a computer program, with the goal of training the program to understand how the world works. For example, how is the camera on our phones able to recognise a human face when it spots one? Well, to answer crudely, by feeding the software millions (or even billions) of labelled data of human faces.
So, as more computing power and data became readily available, AI researchers and their algorithms leaned towards the neural networks approach, employing technical breakthroughs in AI tools, such as Deep Learning, to upend the world as we know it.
AI’s grim reputation as the instigator of technology – robots – that will wipe out the human race is still very much in the realm of science fiction, despite dire warnings from eminent personalities like Elon Musk and Stephen Hawkings. What we should be afraid of, Lee tells us in AI Superpowers, is the impact the technology will have on the global jobs market. He asks a profound question, “When machines can do everything that we can, what does it mean to be human?”
There is good reason, of course, to believe that the efficiencies brought about by the AI revolution will not just take away jobs but create new ones, just as it happened with the steam engine, electricity and the computer revolution. But Lee carefully dispels this notion, noting the breakneck speed at which AI is transforming entire industries. “Whereas the industrial revolution took place across several generations, the AI revolution will have a major impact within one generation,” he writes.
This speed is as a result of many factors: globalization, openness and transparency within the AI academic community and the input of one country that didn’t participate in the industrial revolution: China.
China’s technology scene used to be famous for its copycat syndrome. But that reputation is gradually being scrubbed off as the country enters a new era of upgrading its technology ambitions. And this is important, as the country represents almost one-fifth of the world’s population and is now the second biggest economy in the world. If it is now an active player on the AI scene, then the world’s brainpower has literally multiplied.
Moreover, Lee’s intricate analysis of the Chinese tech world as compared with Silicon Valley, which has dominated the global tech scene for decades, is elegantly striking in its portrayal of China as the perfect soil for the fertile implementation of AI. China, unlike Silicon Valley, is less bogged down by the pursuit of clean, elegant solutions designed to make the world better for everybody. Rather, the country’s tech fabric is driven by what Lee describes as ‘tech-utilitarianism’, an idea that can be described as ‘the application of technology for the greater good’. In other words, one is elitist while the other is down-to-earth; one is idealistic, while the other is realistic, one is obsessed with personal privacy, the other with community gains.
An obvious instance is data use and collection. Whereas Silicon Valley companies are under intense spotlight on how they use the data they collect and their customers are less willing to share their personal data, in China the reverse is relatively the case. And, in AI, the more data you have, the better your algorithms can perform.
Lee sees this Chinese pragmatism as an opportunity for China to catch up – and even surpass – the United States in terms of AI sophistication. And he is not alone. According to PWC, AI implementation is estimated to add $15.7 trillion to global GDP by 2030 and China is predicted to take home $7 trillion of that total, nearly double North America’s $3.7 trillion in gains.
During his TED talk in Vancouver, Kai-Fu Lee shared a couple of personal stories which detailed his transformation from a workaholic to an adult who began to realise that the most important things in life cannot be quantified. For someone who had built a career around the power of computation (in 1988, he used a technique akin to neural networks to create Sphinx, the world’s first speaker-independent program for recognising continuous speech), this was new territory for him. But he embraced it, and when it was time for him to look forward, this new perspective coloured his outlook.
If AI is going to take our jobs, some of the smartest minds have suggested retraining workers for the new jobs that will be created, reducing work hours so more people can fill a spot and income redistribution, a kind of universal basic pay that ensures people can meet their basic needs. But Lee is unconvinced that these approaches will be enough in themselves. The new jobs, he argues, might not be enough, and giving people ‘free’ money strips them of dignity and purpose.
His ideas for the future, instead, come from the recognition of what he believes differentiates man from machine. “For all of AI’s astounding capabilities,” he writes, “the one thing that only humans can provide turns out to also be exactly what is most needed in our lives: love.”
He advocates for service-focused impact investing in venture capital, which will have to accept linear returns in exchange for meaningful job creation and also for a social investment stipend, “a decent government salary given to those who invest their time and energy in those activities that promote a kind, compassionate and creative society.”
Featured photo credit to ted.com