China experienced its sputnik moment in March of 2016. That month, AlphaGo, an artificial intelligence program, defeated a South Korean grandmaster at Go, a challenging Chinese board game. Around 280 million people in China tuned in to watch the five-game series. AlphaGo defeated China’s 19-year-old prodigy Ke Jie in May 2017, and the Chinese government unveiled an AI policy two months later, predicting that the country will become the world’s center of AI research by 2030. After the year, China accounted for more than half of all AI venture capital funding globally.
According to Kai-Fu Lee, a recognized authority on the subject, in his book AI Superpowers, this is how China became an AI powerhouse. Lee, 60, is the founder and CEO of Beijing-based online investment firm Sinovation Ventures, as well as a former Google China president who has witnessed the country’s digital development firsthand. According to the book, which was published three years ago, China would outperform the field’s other major competitor, the United States, in the AI race.
I spoke with Lee this week and inquired if China was winning, as he predicted. He is kind, stating that “both nations are equally powerful,” but that they are developing in different areas: the US in enterprise software (what businesses use), and China in manufacturing and image recognition. Both countries are “equally powerful in aggregate,” but “stronger individually” in various sectors, according to him.
But first, I’d like to include my regulatory AI explainer in case you’re unfamiliar with it (it’s one of those classic pieces of technology that has already become embedded in different aspects of your life).
AI can explain…
As described in Lee’s book, the neural network kind of AI replicates the underlying architecture of the brain in computer form. This includes creating layers of artificial neurons. They carry information between sections of the brain and the rest of our nervous system in humans. Artificial neurons will receive and send data in a way similar to biological neuron networks. These networks are then fed a huge number of examples of a particular phenomenon. For example chess moves, animal images, or noises. They are trained to find patterns in the data.
Lee uses cats as an example in his book: using the neural network technique. The software is given millions of sample photographs labeled ‘cat’ or ‘no cat,’. The program will allow it to find out for itself which aspects in the millions of images are most closely connected to the ‘cat’ label.
“Neural networks need a big number of two things: computational power and data,” argues Lee. The data ‘trains’ the software to recognize patterns by providing it with a large number of instances. The computational power allows the program to interpret those examples at rapid rates. “Deep learning is now referred to as “juiced up neural networks”. He goes on to say that AI algorithms would have an impact on white-collar occupations. Due to their ability to analyze data and make judgments based on it more effectively. AI algorithms is a set of instructions that a computer program follows to accomplish a task or create a particular output –
“A large portion of today’s white-collar workforce is paid to take in and analyze information. Then make a judgment or suggestion based on that information. It is precisely what AI algorithms excel at.”
And now we’re three years later. One former US government official has already weighed in on the AI superpower competition. Nicolas Chaillan, the Pentagon’s first top software officer, quit earlier this year because he couldn’t stomach watching China outpace the United States in artificial intelligence, machine learning, and cyber capabilities. “We will have no fighting chance against China in 15 to 20 years”. “In my perspective, it’s already a done deal; it’s already ended,” he argues, and China is positioned to control everything from media narratives to geopolitics.
Is there no fighting chance?
Lee’s book is a wonderful place to start if you want to learn how China became a digital giant. This year he published another book on AI, co-written with speculative fiction author Chen Qiufan, called AI 2041. It illustrates how, initially, the Chinese tech sector copied US platforms (Twitter, Facebook, eBay) before generating dominating domestic services like WeChat (the app that does everything) and a worldwide sensation like TikTok. These variables account for a large portion of Lee’s optimism about China’s position in the global AI race.
According to Lee, coronavirus has pushed AI adoption in the United States and China. We’re on Zoom, and he points to the hazy screen behind him as an example. “AI created the hazy background.” “How could you tell who’s me and what’s backdrop if you didn’t have AI?”. He claims that the US lockdown has increased the use of AI in the workplace. It highlights the success of Romania-founded but New York-based UiPath. It offers AI-based software that, automates workday chores for organizations like reading papers and filling out forms. “Companies have all this data,” Lee adds, “which turns into a treasure trove of knowledge on which they can help make the organization more efficient”. China generates a large amount of data.
In China, he argues, the pandemic has pushed the use of robots in businesses as well as autonomous vehicles for transporting goods. “Of course, it extends to restaurants, where many Chinese enterprises use robotic waiters”. “Not fancy restaurants like McDonald’s. But… medium to low-cost sit-down restaurants where a robot delivering your lunch would be safer”. “This isn’t a robot that looks like a human. It’s a tray that rolls up to your table.”
According to Lee, language processing and understanding are a big step forward for AI. He goes on to claim that search engines would change if they could “basically answer inquiries, comprehend text and language. Also if they offer highly meaningful responses to inquiries”. He goes on to claim that AI will enable the development of personalized movies to entertain or educate you and that the metaverse. Metaverse is a digital domain in which we may virtually live our social and professional lives. It would be filled not just by avatars of our actual selves, but also by AI avatars. “A major aspect of the metaverse is that it will be populated not only by people but also by digital humans”. He goes on to say that the process of identifying new pharmaceuticals would be modified as well.
We also obsess over the bad. Lee declines to comment on the use of face recognition in China. There has been opposition but says autonomous weaponry, including drones, is a “very major problem”. “How to govern that is an issue that concerns all of us,” he continues. Just as chemists spoke out against chemical weapons and biologists spoke out against biological weapons, I hope governments begin to listen to AI experts. It’s probably difficult to put a halt to it entirely. However, there should be a method to prevent or minimize the most egregious misuse.”
According to Lee, “deep tech,” which comprises major technological developments like artificial intelligence, quantum computing, and blockchain, will be the next wave of investment and entrepreneurship.
“If you were a bright new graduate ten years ago, you were undoubtedly asking yourself, ‘What apps should I build?'” “I’m sure you’re debating whether or not to pursue a Ph.D.”
Meta’s communication problems
WhatsApp, which is owned by Facebook parent Meta, launched a new feature on Monday. It allows users to delete messages after 24 hours. End-to-end encryption will not be available on Meta’s Facebook and Instagram messaging services until at least 2023. However, the NSPCC’s reaction to WhatsApp’s decision – “poorly planned out” – highlights the ongoing friction between Meta and regulators, lawmakers, and campaigners over privacy.