Recently, a 20-page PDF report from McKinsey gave a comprehensive and detailed introduction to the current state of AI development in China. The article analyzes China's AI strength from four perspectives: academic research, algorithms, data, and computing power. The article pointed out that AI is of great importance to China's development, but the key now is talent.
The article concludes with five strategic recommendations for the development of AI in China: 1. Establish a sound data ecosystem; 2. Promote the adoption of AI in traditional industries; 3. Strengthen professional AI talent management; 4. Establish education and training systems for challenges 5. And build moral and legal consensus between Chinese citizens and the global community.
The main contents of the report are as follows:
In March 2016, the AlphaGo computer program beat Lee Sedol, a nine-dan Go player, which immediately sparked discussions around the world. This milestone event proved to the world that machines can think like humans and even do better than humans. Optimists believe that breakthroughs in artificial intelligence technology will greatly boost productivity. But it has also fueled anxiety that artificial intelligence may replace human jobs, and some even worry that humans will eventually create intelligent machines that they can't even control themselves. Beneath the myriad of views, one thing is clear: AI has enormous potential to transform global society.
With the rapid disappearance of the demographic dividend, China urgently needs to find a new growth engine. AI-based automation could boost productivity and help China achieve its economic development goals.
Artificial Intelligence: The inflection point is coming
Artificial intelligence means the simulation of the cognitive functions of the human brain by machines. This concept has long existed only in human fantasy and science fiction, and it was not until the 1950s and 1960s, when theories about artificial intelligence were first formed, that general optimism and the first wave of enthusiasm began. However, due to the failure of technology to achieve breakthrough progress, artificial intelligence could not achieve the expected effect, so it fell into a period of silence. Although there are many success stories in the following decades, the success stories of artificial intelligence in the real world are too isolated to support large-scale commercialization.
Let's fast forward to the 21st century.
Technological advances in data collection and organization, algorithms, machine learning, and high-performance computing have led to revolutionary advances. For example, in the game of Go, which was previously considered unwinnable, AlphaGo successfully defeated the human world champion, thus giving the victory a historic significance.
And change isn't just happening at the theoretical frontier.
Applications in industries such as finance, medical care, and manufacturing have developed rapidly, and global venture capital in the field of artificial intelligence has soared from $589 million in 2012 to more than $5 billion in 2016. McKinsey predicts that by 2025, the market value of artificial intelligence applications will reach 127 billion US dollars.
Understanding AI and what it can do
Traditionally, we have used the processing power of computers to produce outputs more efficiently (eg, perform faster and more complex calculations than humans can). Traditional software programs have been programmed with specific instructions to perform tasks. AI systems take a very different approach. They can unearth patterns, connections, and insights from huge sets of “big data,” and they also employ generalized learning strategies that allow them to adapt to new data inputs without explicit reprogramming. Systems utilizing machine learning have inductive and decision-making capabilities, and the advent of deep learning has pushed the boundaries of this capability even further. Today’s machine learning systems are able to learn, discover and adapt to rules on their own.
While recent breakthroughs in deep learning have produced artificial intelligence systems that can match or exceed human intelligence in certain key functions, we are nowhere near "general AI" - or machines that can perform comprehensive cognitive tasks like humans do some distance. Many machine learning systems have been used for specific commercial purposes, and the applications are very diverse. They can serve customers, manage logistics, monitor equipment, optimize energy consumption, and analyze medical records. Recent research by the McKinsey Global Institute (MGI) shows that machine learning techniques have widespread applications in almost every industry.
To understand the capabilities of AI, it is a good way to look at the following four dimensions:
- perception
- predict
- Guidance method (prescription)
- Comprehensive solutions (combined with technologies such as robotics, autonomous driving, etc.)
The current level of commercialization varies by various AI capabilities. While systems with sensing and predictive capabilities are already on the market, more prescriptive tools and integrated solutions are still being developed (Figure 1).
Figure 1: The current commercialization of AI technology, McKinsey believes that the commercial applications of IBM and iFLYTEK belong to perception technology. On the other hand, Baidu and Amazon are solutions that combine hardware.
The Future of AI: Difficult Challenges and Possibilities
Technological advances in the past have primarily focused on enhancing the ability to perform clearly delineated production tasks. But now, AI enables machines to react and adjust to optimize results. Combining the Internet of Things (IoT) and robotics, it can create an integrated cyber-physical world.
Current trends suggest that AI technology will eventually gain global acceptance in a wider context and industry, and one of the most important outcomes is to handle a variety of tasks that have long been performed by humans. The McKinsey report analyzes more than 2,000 job activities in more than 800 occupations in the global economy. Technically, 50% of work activities can now be automated using the currently demonstrated technology.
But technical feasibility is only one factor that affects the pace and extent of automation. Others include the cost of developing and deploying specific applications, labor market dynamics, economic benefits, and regulatory and social acceptance. Taking these factors into account, McKinsey's research on automation suggests that half of today's work activities will not be automated until 2055, but there is considerable uncertainty over that timing. In the case of active adoption, this level of automation could happen 20 years earlier, and in the case of later adoption, it could happen 20 years later.
Along this line, AI can be a powerful tool for some of society's core challenges. In medicine, AI will greatly enhance our ability to analyze the human genome and develop personalized and more effective treatments for each patient. It can greatly speed up the process of curing cancer, Alzheimer's disease and other diseases. AI systems can analyze weather patterns at scale, improve energy efficiency, and improve our ability to monitor and respond to climate change. The possibilities are beyond our imagination. For example, AI systems could one day open up the exploration of Mars and outer space.
The significance of AI to China: Algorithms, data, computing power and other countries' horizontal comparison
China has become the world's leading AI R&D center as China's big technology companies push forward AI research and development. China's huge population base and diverse industry mix have the potential to generate massive amounts of data and form a huge market. Widespread adoption of AI technologies is critical to China's future economic growth, as the nation's aging population accelerates the need for productivity growth, including a more open data environment and well-trained data science talent. But AI also raises more complex social and economic issues that require careful consideration.
China's place in AI development
China and the United States are currently leaders in global AI development. In 2015 alone, the two countries published close to 10,000 AI-related papers in academic journals, while the United Kingdom, India, Germany and Japan combined accounted for about half of China and the United States. (Data source: SCImago Journal & Country Rank, 2015)
Much of China's AI development has been driven by private high-tech companies. With the help of massive search data and diverse product lines, some of China's Internet giants are leading the way in technologies such as image and speech recognition. And these technologies are already being incorporated into their new products, including smart assistants, self-driving cars, and more.
China has reason to be optimistic about its role in the AI-defined future. China's huge population can generate massive amounts of data, a prerequisite for "training" AI systems. China also has the advantage of "economies of scope": a wide range of industries provides fertile ground for products to be deployed in the market.
However, in order to remain at the forefront of this rapidly developing field, China still needs to spare no effort and maximize the economic potential of these technologies. China needs to focus on enhancing innovation capabilities. For example, while Chinese scholars publish more AI-related papers than US researchers, their papers are not as influential as US and UK researchers (see Table 2).
Table 2: The left graph is ranked by the number of AI-related papers, and the right graph is ranked by H-index. Although China has published a large number of widely cited AI-related papers, the US and the UK are still more influential in terms of influence. China ranks first in absolute citations, but after self-citations are removed, the U.S. has an edge.
In addition, China has not yet formed a vibrant AI ecosystem like the United States, which is reflected in the fact that the United States has much more AI startups than China (see Table 3). The U.S. ecosystem is large, innovative and diverse (including research institutions, universities, and private companies), and it benefits from Silicon Valley's technology industry, with advantages that are difficult to replicate.
Table 3: The AI startup ecosystem in the US is stronger than in China. The figure above shows the proportion of China and the United States in the 50 largest AI startups (ranked by total financing value), and the data comes from the AI 100 list published by CB Insight. via: CB Insights; McKinsey analysis
data
Just as humans get their energy through food, AI cannot operate without a stable source of data. These systems must have vast amounts of data on which they can “train” and continually improve and refine the results they produce. On the data side, there are several issues that could hinder China's AI development.
First, big tech companies in China collect data through their proprietary platforms, and China lags behind the US in creating a data-friendly ecosystem that lacks uniform standards and sharing across platforms. Second, countries around the world have found that open government data helps private sector innovation, but China's public sector has relatively few open data (see Table 4). Finally, restricting data flows across borders also puts China at a disadvantage in global cooperation.
Table 4: Openness of government data, China ranks 93rd in the world.
Explanation: Each data category is assessed based on 10 factors of public accessibility, including whether the data is published online, free, up-to-date, and machine-readable, among others. Source: Open Knowledge International, 2015; McKinsey Global Institute analysis
algorithm
At the application level, China is comparable to other countries in algorithm development. In fact, Chinese researchers have made breakthroughs in developing algorithms for speech recognition and targeted advertising. Thanks to the global open source platform, Chinese companies can quickly replicate the most advanced algorithms developed elsewhere.
However, China lags behind the US and the UK in basic research. A major reason is the shortage of talent, and recruiting talent is critical to China's AI development. More than half of data scientists in the United States have more than 10 years of work experience, while in China, as many as 40% of researchers with less than five years of experience.
China currently has less than 30 university research laboratories focusing on artificial intelligence, and these laboratories alone cannot output enough talents to meet the recruitment needs of China's AI industry. In addition, Chinese AI scientists are focusing more on areas such as computer vision and speech recognition than in other specialized fields. AI programs at universities can also benefit from higher math and statistics requirements, striving to remain a global leader in the field. Also consider changing the mode of providing research funding to promote more innovation.
computing power
Computing power is not a direct bottleneck to the commercial development of artificial intelligence in China. With the widespread use of microprocessors in the global market, computing power has become something that is easily available.
But China still cannot ignore the importance of developing its own advanced semiconductor, microprocessor and high-performance computing technologies. Computing power is one of the foundations of AI and is of strategic importance.
China has historically relied heavily on foreign microchip suppliers. For certain types of high-value semiconductors, China relies almost entirely on imports. However, in 2015, the US government banned the world's three largest chip suppliers Intel, Nvidia and AMD from selling high-end supercomputer chips to the Chinese government. Achieving greater control over the supply of core technologies could help improve China's ability to deploy AI systems more widely in the future.
In order to solve this problem, the Chinese government released two policy documents in 2014, the "National Integrated Circuit Industry Development Promotion Outline" and "Made in China 2025", and the government set up a fund of more than 20 billion US dollars to do this. These initiatives are already beginning to bear some fruit.
Special-purpose processors, such as graphics processing units that can perform a lot of complex calculations, are especially important for AI. With the development of China's IC industry, the development of such processors should also be given enough attention.
In China's strategy on artificial intelligence, it is important to note the increasing globalization of the technology industry. All aspects of the AI value chain, from basic research, to application development, to hardware manufacturing, involve global collaboration. In addition to building its own data ecosystem, data science talent pipeline, and semiconductor industry, China needs to ensure that its AI industry is built on open systems that integrate with global markets.
The economic impact of AI development: 0.8 to 1.4 percentage points of GDP growth in China
AI is an important opportunity for China to accelerate the development of productivity, and it is also a key to solving the aging population. However, policymakers also need to consider and prepare for the potential disruption to labor markets that AI could bring.
In recent decades, China's development has benefited from a "demographic dividend" as an expanding labor force fueled economic growth. But as the population ages, China will lose that advantage. Research shows that the country's working-age population has peaked and will continue to shrink for decades to come. This demographic trend means China will struggle to maintain the workforce needed for economic growth at current levels of productivity. The only option to maintain momentum is to substantially increase productivity growth.
AI can partially close this gap. AI systems can increase productivity by helping or replacing humans to perform existing work activities more efficiently. Intel, for example, collects vast amounts of data in parallel with its chip manufacturing process to make improvements, whereas in the past, the company has relied primarily on human labor to perform root-cause analysis of the data if errors occurred. But now machine learning can accomplish this task much faster than humans, with algorithms sifting through thousands of data points about each chip to find common patterns in those with defects. Additionally, AI can make processes such as industrial machinery, supply chains, logistics routes, and more more efficient. AI applications can create superior efficiencies by predicting failures, identifying project bottlenecks, and automating processes and decisions.
A large part of China's economy includes hotels and food services, manufacturing, agriculture and other sectors. According to the MGI report, AI-led automation could make the Chinese economy more productive, adding between 0.8 and 1.4 percentage points of GDP per year, depending on the pace of adoption.
In addition to boosting productivity, the rise of AI will most likely create new products and services, which in turn will spawn new occupations and businesses. Only a few decades ago, no one could have imagined that there is now a large number of jobs related to the Internet economy, and AI has a similar transformative effect.
AI has the potential to dramatically increase productivity growth, but it could lead to greater income disparities. Fewer and fewer people will be needed in roles such as customer service. Overall, AI will add to the trend of so-called "skill-biased technological change," where there will be a new premium for digital skills, but at the same time, there will be less demand for low- and medium-skilled workers. This may reduce overall labor demand. While average income may rise, polarization will increase. The "digital divide" can manifest itself as a social divide.
Overall, China's workforce can be automated more than any other country in the world. MGI estimates that 51 percent of jobs in China can be automated, equivalent to 394 million full-time employees. However, even with early adoption, around 90% of work activities will be automated, and China may still face a shortage of labor needed to achieve its 4-5% GDP growth target by 2055. This will make China look for more ways to increase productivity.
Routine jobs and predictable, programmable tasks will be particularly vulnerable to replacement by AI. Mid-skilled workers may bear the brunt due to cost-effectiveness, while low-paying positions may last longer. That's not to say, however, that today's high-skilled jobs will be completely immune to disruption. Many tasks performed by professionals with specialized knowledge and experience, such as doctors, may be automated, and these jobs may change to focus more on personal interactions. Many jobs will not disappear, but their mix of activities will change, and education and training systems will need to change accordingly.
A recent U.S. government report identified four categories of AI-related jobs that may become commonplace in the future: Engagement jobs that require working with AI systems to complete complex tasks (such as routine nurse-patient checkups using AI applications) ; development work that creates AI technologies and applications (such as database scientists and software developers); supervisory work that monitors, licenses, or repairs AI systems (such as technicians maintaining AI robots); and work that responds to AI-driven paradigm shifts ( Such as lawyers creating legal frameworks around AI, or city planners creating environments that can accommodate autonomous vehicles).
The impact of AI on society: Careful regulation should be adopted
AI technologies can improve healthcare, the environment, safety, and education, with exciting potential to enhance human well-being. At the same time, it raises complex ethical, legal and security issues as it blurs the lines between the physical, digital and personal realms. Careful regulation should be exercised in introducing AI into society.
Numerous cases have already demonstrated the potential of AI to solve social problems. Artificial intelligence systems can help scientists predict environmental changes; Cornell University, for example, is using this ability to predict habitat changes to protect certain birds. AI also has broad applicability in healthcare. The Dutch government is using it to determine the most effective treatments for certain patient populations and to reduce medical errors through the analysis of digital health records. In the U.S., Las Vegas is using the technology for public health surveillance, using social media tracking to determine the origin of disease outbreaks.
AI systems can also improve the safety and efficiency of public transportation and transportation systems. There is evidence that driverless vehicles can reduce traffic accidents. Alibaba partnered with the Hangzhou Municipal Government to make urban traffic smarter with AI-integrated traffic lights, reducing congestion and increasing traffic flow by 11% in specific areas. AI is also being used to predict energy needs and manage energy consumption. Early examples include Google reducing energy consumption in vast data centers and the U.K. government's need to manage a surge in its power grid system, showing the potential for AI technology to save companies and consumers billions of dollars.
These unprecedented capabilities raise many ethical and legal issues that require serious consideration. Asimov's famous Three Laws of Robotics was the first attempt to formulate the basic principles by which robots interact with humans. But the ethical issues raised by the advent of AI are more subtle and potentially more impactful.
First, in a world where sensors and various AI systems are ubiquitous, businesses are constantly collecting personal data—not just using digital devices, but also through public and personal spaces. In some cases, such as hospitals, this type of personal information is very sensitive. This raises questions about who should own this personal data, how it can be shared, and how it can be protected from the risk of cybersecurity breaches.
Second, AI may unconsciously discriminate when making decisions. Since the "real world" is full of all kinds of racism, sexism, and prejudice, real-world data fed into algorithms also have these characteristics—when machine learning algorithms learn from biased training data, they internalize eliminate these prejudices.
In addition to these ethical considerations, the adoption of AI by society will also have many legal implications. For example, if an accident or even a crime is committed as a result of an AI decision, who is to blame? Who owns the intellectual property created by the AI system? What are the rules for the power of AI? What legal rights and obligations do AI developers have. These and many other issues require thorough debate to create a sound legal and ethical framework.
The geopolitical impact of AI: some countries may face new social unrest
Developments in the field of AI are truly global. Further development will require international cooperation to facilitate broader access to data, algorithms, capital and talent. But as the digitization of the global economy grows, many aspects of global governance remain a vacuum. Many ethical and safety issues posed by automated systems with greater than human intelligence need to be addressed not only at the national level, but also through international cooperation.
Furthermore, just as AI-driven automation could create a two-tiered labor market within individual economies, it could widen the “digital divide” globally, with slower technological progress countries becoming more behind. Some countries with rapidly growing populations and relying on labor-intensive economic development models may even face new social unrest due to mass unemployment.
Finally, computer simulation tools are already widely used in some war games, and AI will further improve the accuracy and power of such simulations. The weaponization of AI is an area of concern. A U.S. Navy report argues that as military robots become more sophisticated, more attention should be paid to the impact of their autonomous decision-making capabilities. More than 1,000 AI and robotics researchers, including Stephen Hawking and Elon Musk, have signed an open letter calling for a ban on AI warfare, warning of the potential for horrific disruption from "autonomous weaponry." AI systems , like nuclear energy and nuclear weapons, may require strong international agreements to ensure their peaceful use and maintain global security.
China's AI strategy: 5 strategic recommendations
Turning today's technological innovation into China's long-term sustainable growth engine requires a well-thought-out strategy. The government should lay a solid foundation, provide inspiring goals for AI development, and stimulate private sector innovation and the application of new technologies. The strategy consists of a strong industrial and economic framework, an educational framework, and a social and international policy framework.
Industrial and Economic Framework
While the development of AI is still in its early stages, it seems unlikely that the technology will follow a linear growth trajectory. The possibility of a quick start is urgently needed to ensure sound industrial policy. Otherwise, China risks favoring incentives, overinvestment, and oversupply, all of which destroy value. While the market will drive the development of AI technology and its applications, the right policy framework can create a healthy environment for growth.
Strategic Priority 1: Build a Robust Data Ecosystem
Abundant data is a key factor in training AI systems, attracting talent, and accelerating innovation. To build a stronger data ecosystem, China can set and implement data standards, open up public data for individual R&D, and encourage the exchange of international data flows.
Standardization is an important precursor to system-wide data sharing and interoperability that will enhance the value of IoT and AI technologies. Given the sheer amount of data that has the potential to be available across the country, China is uniquely positioned to take the lead in ensuring that Chinese data standards are implemented.
For industry-specific data, governments can call on existing regulators to make the necessary rules. In the United States, for example, the Securities and Exchange Commission mandated in 2009 that all public companies must disclose their financial statements in XBRL (Extensible Business Reporting Language) format, thereby ensuring machine-readable public data.
To increase the diversity of available data to support AI development, governments can open up more public datasets and take the lead in establishing some industry-specific datasets. In addition to advancing the AI industry, these initiatives will also have benefits in improving the quality of public services and interpreting new policies. For example, the city government of New York City launched its own open data portal to give citizens access to data on economic development, health, recreation, public services, and more. New York also enacted an open data law in 2012 that requires the government to process machine-readable data and establish APIs (application programming interfaces) that enable software developers to connect directly to government systems and collect data.
Last but not least, the Chinese government will need to consider the value of international data. The MGI study found that cross-border data contributed $2.8 trillion to the global economy in 2014, with a greater impact on growth than trade in goods. Furthermore, the inflow and outflow of data are important because they reflect the thinking, research, technology, talent and best practices of the global economy. Data is the currency of the future. In medical research, for example, it is impossible to realize the full potential of AI without acquiring data from the vast array of clinical data around the world. Too many barriers could hinder Chinese AI companies from developing competitive products in international markets.
Strategic Priority 2: Expand the proportion of AI adoption in traditional industries
In China, realizing the full potential of AI in the economy depends on the practical application of AI systems in traditional enterprises, not just in the application of technology giants. A great deal of value can be unlocked by increasing the productivity of vast production units. But China needs to address several key hurdles.
The first hurdle to overcome involves changing mindsets and creating a sense of urgency that needs to change the way the business operates. According to a McKinsey survey, AI is not yet a strategic priority in more than 40% of companies in traditional Chinese industries. As a result, many of these companies do not yet have the data needed to support future AI adoption. Agribusinesses, for example, rarely consider recording planting schedules or details about the impact of weather on output, but this is exactly the kind of information AI systems can use to uncover valuable insights and utility. By contrast, the UK, US and Japan have built national information systems to capture this data and apply advanced analytics to modern agricultural management.
The second major hurdle is ignorance of technology. As mentioned above, China will need to focus on developing more elite data scientists, especially in regions where AI skills shortages are becoming increasingly apparent. However, there is also a shortage of talent capable of turning AI knowledge into real-world applications with practical value. More business leaders and middle managers need technical skills and the ability to understand and apply data. A Chinese chipmaker like Intel recognized that data generated during manufacturing and testing could significantly improve operations and reduce defective products. But the company was unable to implement the strategy due to a lack of experts in semiconductor and AI knowledge.
Last but not least, AI adoption is impacted by cost. Buying AI systems and hiring the scarce and specialized talent needed to maximize their value is not always cost-effective for Chinese companies. When labor costs are low, there is less urgency to use technology to streamline manual processes.
The greatest economic potential that AI can bring to China is the innovation of traditional industries. Market growth can be boosted if governments can help traditional industries lower the barriers to adopting AI technologies.
To facilitate AI adoption, policymakers should focus on helping the market overcome the three main barriers discussed earlier in this article: lack of strategic awareness, AI adoption costs higher than labor costs, and AI ignorance.
Some of these issues can be addressed through traditional tax credits and subsidies. Governments may also consider adopting AI systems within government agencies. This has a strong follow-up effect, boosting market launch, supporting government suppliers, and ultimately reducing adoption costs by accumulating technical experience and talent.
In addition, encouraging traditional industries to adopt the Internet of Things (IoT) will lay the foundation for more value from AI adoption, as IoT can connect networks of sensors and devices together to provide AI systems with massive amounts of real-world, real-time data. The government can focus on creating some IoT success stories in key economic sectors to complement its “Internet Plus” policy, thereby establishing a model that other traditional industries can follow.
Educational Framework: Talent is critical to the development and adoption of AI. A strong talent pyramid should have top scientists pushing the boundaries of AI-based technology, many developers capable of creating AI applications for real-world environments, and a large workforce capable of working with AI systems in a variety of work environments .
Strategic priority 3: Strengthen the delivery of professional AI talents
To address the AI talent gap in China, the government needs to invest in AI-related education and research programs, reposition the education system with a greater focus on innovation and digital skills, and formulate immigration policies to attract the best global talent.
To develop more elite computer scientists, the technology needs to be advanced, and the government can invest in creating AI projects and fund AI research labs at top universities. This could include establishing artificial intelligence centers at top Chinese universities, or sponsoring innovative research centers that foster collaboration between universities, research institutes and private companies. The South Korean government recently took a big step in this direction by investing 1 trillion won ($863 million) to form a national public-private joint AI research center with South Korea's leading conglomerates. The Canadian government has made a similar move, investing more than $200 million in AI research programs at three Montreal universities.
Many of the experts we interviewed believe that China needs to focus on building a broader culture of innovation in order to achieve AI breakthroughs. One way to address this is to introduce university courses that combine AI with other disciplines. Top U.S. universities, such as Stanford and MIT, have created joint majors that combine computer science and the humanities to develop new worldviews and inspire creativity. Initiatives of this type could inspire new AI applications in the global economy, spanning health care, law, media, and more.
Investing in university projects will pay off in the long term, as talent is an important magnet for international businesses. Large AI developers are increasingly looking to draw talent from academia. Two-thirds of Google DeepMind's researchers are from academic institutions such as UCL, Oxford and Université de Montréal. Top companies will naturally gravitate to cities with large AI talent. For example, in Montreal's growing AI reputation, both Google and Microsoft have responded by investing in the city's university AI labs and expanding their local offices.
In addition to cultivating more local talents, China also needs to cooperate with top data scientists from all over the world and participate in the global cooperation zone. This includes actively recruiting international experts to work in China, and encouraging Chinese AI developers to go out and absorb the latest research results from around the world. This could require the government to relax some residency and immigration rules, as well as provide incentives and support.
Strategic Priority 4: Ensure education and training systems are prepared to develop technical skills and retrain large workforces
While it may take decades for AI to be widely adopted across the economy and society, China needs to prepare for rapid disruption at the industry level. Some jobs may disappear within a few years of key technological breakthroughs. Most of the typists, phone operators, and darkroom film developers have disappeared because technology has made these jobs obsolete.
Helping the workforce in the most affected industries to adapt and acquire more relevant new skills will be a critical ongoing challenge to maintaining public welfare and social stability. Governments will need to proactively identify the jobs most likely to be automated and ensure retraining programmes are offered to workforces whose livelihoods are at risk. These efforts may involve working closely with vocational training schools and providing educational opportunities to workers.
China will also focus on developing workforce skills related to AI in the long run. This includes not only building the pipeline of future data scientists and engineers, but also ensuring that more employees are able to work with AI technology in a variety of business and professional environments. Science, technology, engineering and mathematics must be emphasized in schools; even basic education and vocational courses need to foster data literacy.
Since AI automation of many routine jobs has the potential to widen the digital divide, governments need to regulate the effects of AI automation on inequality. One aspect of this is ensuring equal access to educational opportunities. This includes ensuring that female students and students from rural and inland areas have adequate access to relevant courses in STEM and AI.
Social and International Frameworks: The advent of AI has the potential to profoundly change society. There is a need for consensus on some of the most pressing ethical and legal issues, both domestically and internationally.
Strategic Priority 5: Building ethical and legal consensus among Chinese citizens and the global community
Domestically, reaching consensus requires a transparent and broad consultation process. This is particularly important for the development and adoption of AI in areas of law, such as driverless vehicle privacy protection and liability. The Chinese legislature needs to provide a framework to remove legal uncertainty.
Once the legal framework is in place, the government needs to establish a regulatory body to oversee and regulate AI activities. Since AI technology will be widely used in various industries, it will require professional opinions from multiple industries. In health care, for example, the consequences of irrational adoption of AI technologies can be severe; the National Health and Family Planning Commission needs to have a strong voice in guiding development.
Internationally, China could take the lead in forming a council to promote the peaceful, inclusive and sustainable development of AI technology. The goal of this international body should be to regulate artificial intelligence, establish standards and develop ethical guidelines.
In addition to regulation, China can also start from the perspective of economic development. To ensure that the global digital divide does not become a permanent obstacle to prosperity, China can share its AI technology and expertise with disadvantaged countries, forming an AI Belt and Road Initiative.
AI has the potential to fundamentally shape our society for decades to come. It is a uniquely powerful tool for China to increase its productivity and maintain its growth momentum. In addition, China has the ability and opportunity to lead international cooperation in the development and regulation of AI, ensuring that this breakthrough technology positively contributes to the general welfare of all mankind.