At COSM 2021, Kai-Fu Lee — computer scientist, writer, venture capitalist and former head of Google China — provided a future cast of the five ways artificial intelligence will change the world. Lee’s predictions are compelling because he takes a tempered view of the capabilities of AI.
Lee says some people misunderstand AI. It can’t replicate the human brain because it works differently from the brain. AI is good at using large amounts of data for numerical optimization and individualization, but very poor at extraction analysis, common sense, insight, and creativity. Lee told the gathering:
… of course [AI] has no self-awareness, consciousness, or emotions or love. So, it is actually quite a good complement for human beings because we’re not good at what it’s good at, and it’s not very good at what we’re good at. (COSM 2021, November 11, 03:58 min)
With those limitations in mind, Lee believes that a materials revolution is on the way and that China is poised to lead the way in AI-based technologies.
Natural Language Learning
In the last decade, programmers have developed machine-learning algorithms that recognize objects. With enough data points, algorithms recognize a face in a photograph. The next phase in machine-learning is natural language learning (also called Natural Language Processing), which takes human speech and turns it into something a machine can understand. Speech-to-text programs and devices like Siri are examples of natural language learning. Lee says the future is simultaneous speech translation and better capabilities in advertising to individuals and answering questions.
Lee says that all of these commercial opportunities will eventually lead to “treating texts not as data, not as information, but as knowledge and insight. And it will open up many applications.” [COSM 2021, (06:11)]
Transportation: Autonomous vehicles
Lee is optimistic that truly self-driving cars will one day happen, but he is also realistic about the practical requirements. Such autonomous vehicles must be rolled out step-by-step, starting in simple environments — apart from people — where the vehicle’s AI system can acquire more and more data before it is ready for widespread use.
What will it take for cars to be truly autonomous? Here’s Lee’s answer: “I believe firmly that this is going to require modification to our existing highway and… city infrastructure. For example, in some cities in China, pedestrians and cars are being separated. The purpose is simply to minimize pedestrians getting hit by cars, thereby making autonomous vehicles launchable.” (07:32)
The more data the system collects, the better autonomous vehicles will be at navigating complex environments. He sees automated vehicles as providing the core of the next operating system because, he says, when you have autonomous vehicles, you’ve got all the capabilities that can also power the future of robotics. (08:17)
Lee is particularly interested in how AI can change healthcare. From genetic tests to wearables to blood tests, healthcare is becoming increasingly data driven. Not only is it data driven, but this data comes from multiple sources. AI can combine data from wearables, health records, genetic sequencing, multi-omics, and blood tests to give a more complete picture of a person’s health, which would aid doctors in early detection of diseases.
Compellingly, Lee says AI could be used to “treat” longevity in the same way that it is used to treat diseases. By comparing a person’s biometric and demographic data to similar people, an AI could recommend an optimal health regime for promoting longevity. He has tried this himself:
I can tell you that the AI feedback, of course filtered by a human and given to me about my sleep eating exercise, as well as nutrients and medicines that I take, has made my blood look six years younger in the past one year.” (10:58)
Lee doesn’t forecast that AI will replace doctors and it will require adaptation for the healthcare context. For example, an algorithm may recommend a treatment based on a utilitarian cost-benefit calculation that violates the norms of medical practice:
So if we save 5% people’s lives and 3% die in the process, AI people would say, great, gather more data, do more of that. We’re saving lives in aggregate, but in medicine with Hippocrates Oath, how can you possibly hurt the 3% of people who would not have been hurt with humans, uh, at, at the helm?” (10:14)
The last two areas that Lee predicts AI will fundamentally change are also the two areas where China will likely lead the way, if for no other reason, than out of necessity. The first is robotics; the other is energy.
China is often dubbed the “world’s factory,” but as the cost of labor continues to increase, he notes, the country will need to automate some of its manufacturing processes. He predicts that factories will automate inspection, movement, manipulation, and delicate operations, such as grasping and flying. He also predicts that 3D printing will factor into advancements in robotics, among other industries, and that eventually robots will be self-maintaining.
Lee’s last prediction, at first, does not seem to be relevant to AI. He predicts that solar energies and batteries, will continue to become more efficient and less expensive And China’s industries lead in solar panels and lithium batteries. Here’s where AI comes in: As costs decrease and efficiency increases, energy consumption will go from relying on natural resources to manufactured sources. According to Lee, manufacturing-driven energy sources will be automated using AI. (13:37)
Why does Lee think that China will be the Leader in AI?
In his book, AI Superpowers: China, Silicon Valley, and the New World Order (2018), Lee forecasts that the more data on which AI can operate, the better the algorithm performs and compared to any other country, China has the corner on data just by way of its population size. He believes that the Chinese people also have the drive and work ethic necessary to dominate the field of AI.
Lee’s vision is of a materials revolution in which goods becomes cheap because labor is automated, materials are made through 3D printing, and energy sources no longer consume natural resources. This materials revolution has the potential to alleviate poverty and the worries of day-to-day living, which will allow people to do what only human beings can do:
People can have a comfortable life, and we’ll have to think about the work ethic that people may, may shift [to]. And also, the opportunity we have as we are liberated from routine work; we can go towards aspirational and creative pursuits. (15:03)
Update: Session Video
You may also wish to read: Chinese technocracy surges ahead with AI surveillance. So what do the reservations expressed, about “the soul” and “love,” really mean? Both big tech entrepreneurs Kai-Fu Lee and Jack Ma seem to believe in souls but do not believe that souls can be trusted with freedom, the way governments can.
Peter Thiel: Artificial General Intelligence isn’t happening. That whole transhumanist movement is slowing down, he told COSM 2021. But, he adds, What IS happening should sober us up a lot. There’s no road to computers that think like people that wouldn’t take us through 24/7 computer surveillance first. Thiel says. Is that what we want?