Author believes that skillful application of AI will be China’s greatest opportunity to catch up with – and possibly surpass – the United States. Use of AI can also represent an opportunity for all people to rediscover again what it is that makes us human.
Field of AI is trying to recreate human intelligence in machine. Clarity of goal and complexity of the task. Around 1950s great minds of Marvin Minsky, John McCarthy and Herbert Simon worked on that task. Development of AI is cyclical movement with its times of high followed by times of its decline. Development went into two direction: one was rule-based approach and second was approach of neural networks.
Rule based approach is sometimes called also symbolic systems or expert’s systems. Researcher tried to teach machines to think by encoding a series of logical rules. System is good for simple tasks but falls apart when multiple choices are provided.
Neural network camp tried approach that simulated human brains by feeding system with large amount of inputs and letting systems find patterns them self. In order for neural network approach to gain full potential two factors are needed. Large computing power and large data sets. Both of them have grown enormously since start of AI development. Main break into development of self-learning of many different layers of neural network came with Geoffery Hinton in mid-2000s. Approach was labeled deep learning. With this break AI potential opens in area like:
- Human speech
- Image recognition
- Predicting consumer behavior
- Identification of fraud
- Making of lending decision
- Vision function of robots
- Self-driving cars
Deep learning is still »narrow AI« – intelligence that take data from one specific domain and applies it to optimize on specific outcome. Its core power is its ability to recognize pattern, optimize for specific outcome and make a decision. Andrew NG has compared AI to electricity. A breakthrough technology on its own and one that once harnessed can be applied to revolutionizing dozens of different industries.
In order to understand development of AI and which countries will have a leading role in it, it is important to understand in which phase of development we are right now. Much of the discovery phase is done. It would be hard to see some major breakthrough in research. Now we are entering phase of implementation and moment when algorithms need to be turn into sustainable business. That is because once computing power and engineering talent reach a certain threshold, the quantity of data becomes decisive in determining the overall power and accuracy of algorithm. In phase of discovery Western Countries held lead in research, mainly because their ability to attract best talent from around the world matched with good funding, culture and business environment. But it will be China that will benefit most from their usage in implementation phase. Because in transition from age of expertise to age of data China can utilize its size and focus. So by moving from discovery to implementation and from expertise to data, China weakness of finding out-of-the-box approach and need for global elite research don’t play such an important role any more.
AI needs four analogous inputs:
- Abundant data
- AI scientists
- Hungry entrepreneurs
- AI-friendly policy environment
China entrepreneurs are exposed to harder competitive environment then Silicon Valley. They live in a world where speed is essential, copying is accepted practice and competitors will stop at nothing to win a new market. In this environment companies and apps like WeChat are stepping further then their western competitors in usage of data and developing new business models with new ideas every day. Support of China government and its focus to develop China into most AI developed country is additional argument for China AI leadership. Governments support of funding and mobilizing resources to achieve this goal is enormous. PWC estimates that AI deployment will bring 15.7 trillion USD to global GDP by 2030 and China will take 7 trillion out of this.
Implementation of AI will bring two challenges – job disappearance and increased inequality. Some jobs will disappear – accountants, assembly line workers, warehouse operators, stock analyst, quality control inspectors, truckers, paralegals. Issue is that because of speed of AI advancement, industry will not be able to react as quick as before and replace them with new job roles that will come with AI adoption. Rising in tandem with unemployment will be astronomical wealth in the hands of the new AI tycoons. AI tends to winner-take-all economics. More data creates better products that attract more customers that create more data. That is AI circle that creates winner of winners. But AI adoption will bring another more personal challenge – a psychological loss of one’s purpose. A lot of people build their sense of self-worth from the act of daily work.
China technological industry is full of copying. One of the main players in this area is Wang Xing (Wang Shing). They called him The Cloner. He cloned America’s most successful start ups and adjust them to Chinese users. He did that in 2003, 2005, 2007 and 2010. He copies Facebook, Twitter, Groupon (his Meiutan is more successful than original).
Westerns believe that copycat approach is keeping China technological development, since talent and original ideas could not be developed. But in reality, it is a second phase after copy part, that is fierce competition on a local Chinese market and ability to adapt successful outside applications to China reality, that is important and is defining who will be winner on Chinese market.
If Silicon Valley startups are being mission-driven, Chinese companies are first and foremost market-driven. This culture of copying is coming from Chinese culture, where memorizing old text was conditions to become member of imperial bureaucracy. And focusing on success more then on fame and glory comes from long history of poverty and conditions where Chinese entrepreneurs are product of one child family with all the hopes of their parents and grandparents being put in him. Responsibility to earn money was major driver for their whole life. Copycat culture upgraded then with own works after copycats understood principles and work behind original works was seen already in 17. and 18. Century when Chinese masters from Guangzhou (then Canton) was first copying clocks that were delivered in China as presents from Europeans to emperor and then started to create their own ones. Since China didn’t have necessary infrastructure to support innovations, copying was a necessary steppingstone on the way to more original and locally tailored technology products. Original products don’t tend to adjust them self for different user preferences, but products coming out of copycat culture are all about adaptivity. One size fit all approach of original was good for starting era of internet, when most countries lagged so far behind that no local version was possible.
American companies like e-Bay, Google, Uber, Airbnb, LinkedIn, Amazon all tried and fail in China. They simply couldn’t adjust to fast past of changes needed in this market to address consumer preferences that are different in China then in USA. They were not able to attract best talents, since local workforce didn’t perceive those companies as an option for them to gain worldwide recognition, local country manager was always seen as platform they could reach.
China community of startups drive its value from highly competitive environment, insane pace of work and it is really market driven, it can utilize lean approach that preaches introducing product on a market as soon as you can and work on it in parallel with customers feedback. American companies that are mission driven and aren’t so agile in everyday fight for survival are more rigid in their approach. China entrepreneurs are using those advantages to win market shares and utilize value of growing consumer-driven economy.
China entrepreneurs that embrace gladiator approach and push their companies from copycat phase into developed phase are:
- Zhou Hongyi – owner of many startups, involved into 3Q war with Tencent with its company Quihoo 360
- Wang Xing – after he sold his Xianonei (Facebook copycat), site reemerged as Renren and in 2008 they went into fight with Kaixin001, that lost a war because his owner wasn’t used of gladiator fight. With his experience he was able to win group-buying war in China with his company Meituan. He was able to win because he didn’t engage in a war, he develop product and didn’t burn cash until war was over and he could capitalize. Today its Meiutan Dianping is valued at 30 billion USD and it is fourth most valuable startup in the world.
Moving into AI era, China try to set up environment that would offer country’s best companies with environment that would resemble Silicon Valley in infrastructure and spirit. Guo Hong was responsible for influential Zhongguancun – technology zone in Beijing. He wanted to create full street inside Zhonggancun where he would put, VC firms, startups, incubators and service providers. In process of building such environment the ecosystem was becoming independent and self-sustaining. It led to an overnight boom in production of the natural resources of the AI age (data). After Guo’s mission, spread of government support for innovation in China was huge. Under the banner of »Mass Innovation and Mass Entrepreneurship« Chinese mayors flooded their cities with new innovation zones, incubators, government backed VC funds. And in order for digitalize world solutions to work, logistic details demanded army of workers to deliver services. In China companies understand that and they get their hand dirty in brick and mortar environment too and that was another advantage to some of Silicon Valley based companies. This ability to also collect data from real world and not only through consumer digital footprint is another advantage of China in AI world fueled with data.
Badui, Alibaba and Tencent are BAT, Chinese strong companies in digital field. China internet usage was different from West since internet was in the beginning made for desktops and laptops, but they were too expensive for Chinese users. But when mobile use started to growth, Chinese user leapfrog computer approach and went straight to cheap smartphones in order to connect them self to internet. Tencent use this opportunity to launch their WeChat application in January 201. They already own QQ instant messaging app and Q-Zone social network, but it was WeChat that became center of development for many areas of digital usage for Tencent. One area was mobile payment, where they clash with Alibaba and their Alipay. Alibaba, Tencent and other Chinese startup started to develop solutions for payment with your mobile devices and by doing so connecting online with offline, by virtually bringing mobile approach to all parts of everyday life.
Important element of Chinese success in the new technologies is government support. Under the slogan »mass innovation and mass entrepreneurship« that government issue mayors around the country set up infrastructure for startups and with help of government financial investments that are setup in a way that they also attract additional VC capital, since majority of upside stays with VC, the whole industry started to grow exponentially.
Explosion of the real-world internet is called O2O – online to offline or turning online action into offline service. Didi Chuxing is company that drove Uber out of China and is now developing even further with new markets and new services. Dianping is an app that started as review site for restaurants and then expended with group buying activities. Tujia is rental site. They are all China businesses that started as digital but went heavy into offline world too. And that is the biggest difference between US startup »light« approach and Chinese »heavy« approach. Diaping offers also delivery, Didi owns repair shops and gas station, Tujia own some of properties. All these heavy movers from online into offline world are grabbing even more data from consumers, by controlling larger part of whole consumer experience. The biggest field where this online to offline approach worked was payment industry with inside app payment mechanism introduced in form of Alipay and WeChat. It enables also peer-to-peer payment and even small transactions, since no charges were applied. In 2017 mobile payment in China surpassed 17 trillion USD. Another field where online activity is changing offline world is bike sharing. By using QR codes for unlocking and paying for bikes this process become quick and easy and companies like Mobike use that to bring bycles back to urban areas of China.
Need for AI expertise is one area where China needed to invest in order to capitalize all the advantages described so far. It did that also with support of government. China is building their army of well-trained researchers and engineers. They are working on quantity over quality, but bigger teams of solid engineers will outperform few megastars researchers in Western world. Some of the biggest companies in AI domain like Google, Amazon, Facebook, Microsoft, Alibaba, Tencent and Baidu, were able to attract some of the biggest talent in AI world and Google is probably most equipped to lead corporate development in AI, but AI is developed also in academic sphere and through startup ecosystem and both of them are more open. Approach of corporate world to AI is more »grid«, trying to commoditize AT. It aims to turn the power of machine learning into standard service that can be purchased by any company – infrastructure used for access to this standardized service is cloud computing and companies behind these platforms like Google, Amazon, Alibaba at as utility companies, managing the grid and collecting fees. Startups are taking »battery« approach to AI, building highly specialized »battery powered« solutions for specific tasks.
In USA Obama’s administration prepare a document of how America can utilize power of AI and released it in 2016. By increasing funding for research, stepping up military-civilian cooperation and making investments into mitigating social disruption America can benefit from development of AI. But report was soon forgotten. In China similar report was released in 2017 and all branches of government with ambitious mayors being in the forefront of actions are working towards achieving goals set in the report. Main cities for startup developments in China are Nanjing, Beijing, Shenzen and Hangzhou. China techno-utilitarian approach with top-down implementation has certain advantage in AI utilization and growth.
AI revolution will sweep world in four waves:
- Internet AI – it is largely about using AI algorithms as recommendation engines, setting up preference-based offerings and suggestions. It was first phase of AI implementation, started almost 15 years ago. USA and China are now on 50-50 ratio in this area, with China having better potentials. One of major AI-internet driven companies in China is Jinri Toutiao, news side driven by algorithms.
- Business AI – is a field where companies are utilizing their data to find correlations of certain strong features and certain outcomes. Optimization like this work well in industries with large amount of structured data on meaningful business outcomes. USA has advantage here since companies are more used of usage of standardized ERP and other applications that produce structured data. China advantage in this field could be lack of standardized systems since new technologies can jump over legacy systems and usage of AI could develop directly new business processes and models. Smart Finance company that offers small loans based on algorithm approval is one of companies that used that advantage. RXThinking is another. Using algorithms to make better medical diagnostic. iFlyTek is helping judges to make better decisions.
- Perception AI – this wave gives machines eyes and ears. It is about digitizing the world around us through proliferation of sensors and smart devices. This new wave is bringing OMO (online-merge-offline). Industries like retail can benefit from those new technologies. Education can benefit. Today all students are forced to learn at the same speed, in the same way, at the same pace and at the same time. Schools take an »assembly line« approach, passing children from grade to grade each year, largely irrespective of whether or not they absorbed what was taught. In-class teaching, homework and drills, tests and grading and customized tutoring are scenarios where AI could be used in order to improve education experience. Services like VIPKid have grown in this field. China has advantage in this area because China is more prepared to sacrifice some privacy for convenience. And balance between personal privacy and public data will benefit China more than USA or Europe. Shenzen is becoming main hub for hardware manufacturing. One of companies that use this Made-in- Shenzen advantage is Xiaomi, company that grow from cheap smartphone producer to producer of network of AI-empowered home devices.
- Autonomous AI – represents the integration and culmination of the three preceding waves. Robots today are already automated, but they are not autonomous. But giving machines the power of sight, the sense of touch and the ability to optimize from data, they can perform almost any task. China could actually benefit from their approach that anything is possible and start optimizing infrastructure to new technologies and self-managed devices. Xiong’an is one of those projects, city build to accommodate new technologies.
Approach to global markets is very different from US companies, that tend to push their own product with as less as possible modification to local markets on other hands Chinese are putting money in local startups and support them in their quest for improving products as much as they can.
AGI – artificial general intelligence – thinking machines with the ability to perform any intellectual task as a human can. Artificial superintelligence or »the singularity« is still a long way out. In order to achieve it we would need a to remove constraints of »narrow AI« and enable new capabilities like multidomain learning, domain-independent learning, natural language understanding, commonsense reasoning, planning and learning from small number of examples. But if AGI is not so close as we might believe, reality of bigger inequality and job disappearing is.
There are those who are afraid of AI development because of above mentioned reasons and those who believe that development will be the same as before, that jobs will be replaced by new ones and overall society will be better off, because of rise of productivity and lower prices. One questions that will stay open is if technology will affect jobs and needed for human workforce will disappear how people will deal with a question: what it means to be human in the age of machines. Will AI become GPT – general purpose technology, one that will disrupt whole industries and will influence whole society – similar like electricity, steam engine and ICT? It will influence cognitive and manual work and speed things up dramatically. And AI revolution will be much faster than others before. Two main reasons for AI fast adoption are fact that algorithms are digital assets and as such quickly deployed anywhere and highly transferable and second is that existence and expansion of venture capital financing. And you can also count China and its power of AI adoption as factor of its worldwide quick expansion.
With job disappearance as one of major threat of AI development, there were different studies with very different estimation about percentage of job lost. Studies focus mainly on America and numbers range from 9-47 percent. Author agrees with higher numbers (but believes that outside factor will reduce actual number to around 10-20%) since not only that AI will cause some job disappearance of already existing jobs because they will be done by algorithms, but they are also startups that are looking for new ways to satisfy the fundamental human needs and they will reshape business environment all together.
When looking for future development of AI, we should have in mind Moravec Paradox – AI is great at thinking, but robots are bad at moving their fingers.
Author believes that the biggest question about AI disruption will be situation where people value that is now in a lot of cases tied to economic activities will be questionable since majority of tasks will be taken by AI tools. He believes that AI approach in optimizing outputs is robbing us of humanity. He lived this way until he was diagnosed with cancer. But after this experience he took on different approach of focusing more on relationship and giving love to the ones around him and less on optimizing his activities in a way to increase his influence as much as he can. Bronnie Ware, palliative care nurse said that in the end it all comes down to relationship and love. Ideal match for future is AI ability to think and human ability to love. Author is proposing that we rewrite our fundamental social contract and restructure economic incentives to reward socially productive activities in the same way that the industrial economy rewarded economically productive activities.
In order to adapt to AI economy realities three R’s are proposed, mainly from influencers from Silicon Valley:
- Reduce – reduce working hours, our let more people work on same job, which could lead to lower wages, but it will decrease full unemployment.
- Retrain – retrain workers that will have access to online training and lifelong learning, so that they can retrain for new jobs that will come out of AI development or that will be outside AI realm.
- Redistribute – but in some cases above will not be enough and some redistribution of income will be needed, and idea of universal basic income can become more interesting.
But author is proposing different solution, one that will encourage private sector to support pro-social activities. That will introduce new jobs or adapt old jobs that will upgrade AI thinking capabilities with human compassion. Some of those jobs already exists but they fall on a margin of economic viability and job stability. Author named them compassionate caregivers. He also believes that creating new ecosystem where VC funds will not chase only returns but will look to support potential service companies that will systematically build caregiving services. Money for these funds will come from government that will look for new way to create jobs and corporation with higher social responsibility goals.
His response that he collectively names social investment stipend – as response to UBI proposal, but this would be granted to people that will invest their time and energy to those activities that promote kind, compassionate and creative society – and those activities would be:
- Care work
- Community services
Author believes that AI productivity will enable such wealth generation that those programs could be financed and that this environment where machines can be used for repeated tasks and also for thinking activities, but people can use this free time for activities that will bring more compassion and love to the ones around them. He wants that we let machines be machines and humans be humans. And that we should choose to use our machines and to love one another.