Editor’s Note:
General Secretary Xi Jinping profoundly pointed out during the ninth collective study of the 19th Central Politburo that accelerating the development of a new generation of artificial intelligence is a strategic issue concerning whether our country can seize the opportunities of a new round of technological revolution and industrial transformation. Accelerating the development of artificial intelligence is an important strategic tool for us to gain the initiative in global technological competition and is a vital strategic resource for promoting our country’s leap in scientific development, industrial optimization, and overall productivity. So, what is artificial intelligence? What historical process has artificial intelligence undergone? How can we accurately grasp the current status of AI technology and industry development? What trends and characteristics will emerge in the foreseeable future for AI development? What is the state of AI development in our country? Please see below for details↓↓
The History, Current Status, and Future of Artificial Intelligence
Tan Tieniu
On February 25, 2018, during the closing ceremony of the Pyeongchang Winter Olympics, an intelligent mobile robot developed by Shenyang Xinsong Robot Automation Co., Ltd. performed with roller skating actors. Photo by Li Gang/Xinhua News Agency
On May 3, 2018, the Chinese Academy of Sciences released the country’s first cloud-based artificial intelligence chip, achieving a theoretical peak speed of 128 trillion fixed-point operations per second, reaching a world-class advanced level. Photo by Jin Liwang/Xinhua News Agency
In October 2017, at the “Future Investment Initiative” conference held in Riyadh, Saudi Arabia, the robot Sophia was granted Saudi citizenship, becoming the first robot in the world to receive citizenship. The photo shows Sophia appearing on the main stage at the Hong Kong Convention and Exhibition Centre on July 10, 2018. ISAAC LAWRENCE/Visual China
On November 22, 2018, at the “Great Changes – Celebrating the 40th Anniversary of Reform and Opening Up Large Exhibition”, the third-generation domestically produced orthopedic surgical robot “Tianji” was simulating surgery. It is the first orthopedic robot in the world to cover surgical indications for the entire spine and pelvic acetabular surgery, with performance indicators reaching an internationally leading level. Mai Tian/Visual China
Just as the steam engine of the steam age, the generator of the electrical age, and the computer and internet of the information age, artificial intelligence is becoming the decisive force driving humanity into the intelligent age. The global industry fully recognizes the significant implications of AI technology leading a new round of industrial transformation, and various sectors are transforming and developing, seizing opportunities to layout AI innovation ecosystems. Major developed countries around the world have made the development of artificial intelligence a major strategy for enhancing national competitiveness and maintaining national security, striving to master the dominant position in international technological competition. General Secretary Xi Jinping profoundly pointed out during the ninth collective study of the 19th Central Politburo that accelerating the development of a new generation of artificial intelligence is a strategic issue concerning whether our country can seize the opportunities of a new round of technological revolution and industrial transformation. Missing an opportunity could mean missing an entire era. The new round of technological revolution and industrial transformation is already in sight; in this major competition concerning our future and destiny, we must seize the opportunity, strive to catch up, and seek to surpass.
Concept and History
To understand where artificial intelligence is headed, we must first know where it came from. In the summer of 1956, scientists such as McCarthy and Minsky held a meeting at Dartmouth College in the United States to discuss “how to simulate human intelligence with machines” and first proposed the concept of “artificial intelligence (AI)”. This marked the birth of the discipline of artificial intelligence.
Artificial intelligence is a new technical science that studies and develops theories, methods, technologies, and application systems that can simulate, extend, and enhance human intelligence. The goal of the research is to enable intelligent machines to listen (speech recognition, machine translation, etc.), see (image recognition, text recognition, etc.), speak (speech synthesis, human-machine dialogue, etc.), think (human-machine games, theorem proving, etc.), learn (machine learning, knowledge representation, etc.), and act (robots, self-driving cars, etc.).
The exploration of artificial intelligence is full of unknowns and has a tortuous path. How to describe the development process of artificial intelligence over more than 60 years since 1956 is a matter of differing opinions in academia. We divide the development of artificial intelligence into the following six stages:
First is the initial development period: 1956 to the early 1960s. After the concept of artificial intelligence was proposed, a series of remarkable research achievements were made, such as machine theorem proving and checkers programs, marking the first peak in the development of artificial intelligence.
The second is the reflective development period: 1960s to early 1970s. The groundbreaking progress made in the early stages of artificial intelligence greatly raised people’s expectations, leading to attempts at more challenging tasks and the proposal of some unrealistic research and development goals. However, a series of failures and unmet expectations (for example, machines being unable to prove that the sum of two continuous functions is still continuous, and machine translation causing misunderstandings) led to a downturn in artificial intelligence development.
The third is the application development period: early 1970s to mid-1980s. The emergence of expert systems in the 1970s, which simulated the knowledge and experience of human experts to solve specific problems, marked a significant breakthrough in moving artificial intelligence from theoretical research to practical application and from general reasoning strategies to the use of specialized knowledge. Expert systems achieved success in fields such as medicine, chemistry, and geology, pushing artificial intelligence into a new peak of application development.
The fourth is the downturn development period: mid-1980s to mid-1990s. As the scale of artificial intelligence applications continued to expand, problems such as the narrow application fields of expert systems, lack of common sense knowledge, difficulties in knowledge acquisition, singular reasoning methods, lack of distributed functionality, and difficulties in compatibility with existing databases gradually became apparent.
The fifth is the steady development period: mid-1990s to 2010. The development of network technology, especially internet technology, accelerated innovation research in artificial intelligence, leading to further practical applications of AI technology. In 1997, IBM’s Deep Blue supercomputer defeated world chess champion Garry Kasparov, and in 2008, IBM proposed the concept of a “smart earth”. These are all landmark events of this period.
The sixth is the flourishing development period: 2011 to the present. With the development of big data, cloud computing, the internet, and the internet of things, information technologies have rapidly advanced AI technologies represented by deep neural networks, significantly bridging the “technological gap” between science and application. Technologies such as image classification, speech recognition, knowledge question answering, human-machine games, and self-driving have achieved breakthroughs from “not usable, not easy to use” to “usable,” ushering in a new peak of explosive growth.
Current Status and Impact
There is some “hype” in society regarding the current status of artificial intelligence development. For example, some believe that AI systems are about to completely surpass human intelligence, that robots will dominate the world within 30 years, and that humans will become slaves to AI, etc. Such intentional or unintentional “hype” and misconceptions can adversely affect the development of artificial intelligence. Therefore, when formulating strategies, guidelines, and policies for AI development, it is essential to accurately grasp the current status of AI technology and industry development.
Specialized artificial intelligence has made significant breakthroughs. In terms of applicability, artificial intelligence can generally be divided into specialized AI and general AI. Specialized AI systems aimed at specific tasks (such as playing Go) can surpass human intelligence in individual tests of local intelligence due to their single task focus, clear requirements, well-defined application boundaries, rich domain knowledge, and relatively simple modeling. Recent advancements in AI have primarily concentrated in the specialized intelligence domain. For example, AlphaGo defeated human champions in Go competitions, AI programs have reached human-level performance in large-scale image and facial recognition, and AI systems have achieved diagnostic capabilities for skin cancer comparable to professional doctors.
General artificial intelligence is still in its infancy. The human brain is a general intelligence system capable of making inferences, integrating knowledge, and handling various issues related to vision, hearing, judgment, reasoning, learning, thinking, planning, and design. A truly complete AI system should be a general intelligence system. Currently, while breakthroughs have been made in specialized AI, research and application in the field of general AI still have a long way to go, and the overall development level of AI remains in its infancy. Today’s AI systems have made significant progress in “shallow intelligence” areas such as information perception and machine learning, but their capabilities in “deep intelligence” areas like conceptual abstraction and reasoning decision-making are still quite weak. Overall, current AI systems can be described as having intelligence without wisdom, having IQ without EQ, being capable of calculations but not “strategizing,” having specialized talents but lacking general capabilities. Therefore, AI still has significant limitations and is far from human wisdom.
AI innovation and entrepreneurship are thriving. The global industry fully recognizes the significant implications of AI technology leading a new round of industrial transformation and has been adjusting development strategies accordingly. For example, Google clearly shifted its development strategy from “mobile-first” to “AI-first” at its 2017 annual developer conference, and Microsoft included AI as part of its corporate vision for the first time in its 2017 fiscal year report. The AI field is at the forefront of innovation and entrepreneurship. A report by McKinsey indicated that global investment in AI research and development exceeded $30 billion in 2016 and is in a phase of rapid growth; another report by global venture capital research firm CB Insights showed that 1,100 new AI startups were established globally in 2017, with the AI sector receiving $15.2 billion in investment, a year-on-year increase of 141%.
Innovative ecological layout has become a strategic high ground for the development of the AI industry. The history of information technology and industry development is a history of the replacement of old and new information industry giants competing for layout in the information industry’s innovative ecosystem. For example, traditional information industry representative companies include Microsoft, Intel, IBM, and Oracle, while companies representing the internet and mobile internet era include Google, Apple, Facebook, Amazon, Alibaba, Tencent, and Baidu. The AI innovation ecosystem includes vertical data platforms, open-source algorithms, computing chips, foundational software, graphics processors, and other technological ecosystems, as well as horizontal smart manufacturing, smart healthcare, smart security, smart retail, smart home, and other commercial and application ecosystems. Currently, the information industry landscape in the intelligent technology era has not formed a monopoly, so global technology industry giants are actively promoting the research and development layout of AI technology ecosystems, striving to seize the high ground of AI-related industries.
The social impact of AI is becoming increasingly prominent. On one hand, AI, as the core force of a new round of technological revolution and industrial transformation, is driving the upgrade of traditional industries and rapidly developing the “unmanned economy,” producing positive impacts in people’s livelihoods such as smart transportation, smart homes, and smart healthcare. On the other hand, issues such as personal information and privacy protection, intellectual property rights for AI-generated content, potential discrimination and bias in AI systems, traffic regulations for autonomous driving systems, and the ethics of brain-computer interfaces and human-machine symbiosis have already emerged, necessitating prompt solutions.
Trends and Outlook
After more than 60 years of development, artificial intelligence has made significant breakthroughs in algorithms, computing power, and data, and is at a technological turning point from “not usable” to “usable,” but still faces many bottlenecks before it can be considered “very usable.” So, what trends and characteristics will emerge in the foreseeable future for AI development?
Development from specialized intelligence to general intelligence. How to achieve a leap from specialized artificial intelligence to general artificial intelligence is both an inevitable trend for the next generation of AI development and a significant challenge in research and application. In October 2016, the U.S. National Science and Technology Council released the “National Artificial Intelligence Research and Development Strategic Plan,” emphasizing the need to focus on researching general artificial intelligence in the U.S. AI mid- to long-term development strategy. Demis Hassabis, the founder of the AlphaGo system development team, proposed the goal of advancing toward “creating a general artificial intelligence that can solve all the world’s problems.” Microsoft established a general AI lab in 2017, with many scientists participating in fields such as perception, learning, reasoning, and natural language understanding.
Development from artificial intelligence to human-machine hybrid intelligence. Drawing on research findings from brain science and cognitive science is an important research direction for AI. Human-machine hybrid intelligence aims to incorporate human roles or cognitive models into AI systems to enhance their performance, making AI a natural extension and expansion of human intelligence, and enabling more efficient solutions to complex problems through human-machine collaboration. Both the new generation of AI planning in China and the U.S. brain initiative consider human-machine hybrid intelligence as an important research and development direction.
Development from “artificial + intelligence” to autonomous intelligent systems. Currently, much research in the AI field focuses on deep learning, but the limitation of deep learning is that it requires substantial human intervention, such as manually designing deep neural network models, setting application scenarios, collecting and labeling large amounts of training data, and users needing to manually adapt intelligent systems, which is very time-consuming and labor-intensive. Therefore, researchers are beginning to focus on autonomous intelligent methods that reduce human intervention and enhance machine intelligence’s ability to learn independently from the environment. For example, the subsequent version of the AlphaGo system, AlphaZero, achieves “general board game AI” through self-play reinforcement learning from scratch in Go, chess, and shogi. In the area of automated design of AI systems, Google’s AutoML proposed in 2017 attempts to reduce personnel costs by automatically creating machine learning systems.
AI will accelerate cross-disciplinary integration with other fields. AI is itself a comprehensive and highly interdisciplinary frontier discipline, with a wide and complex range of research areas. Its development requires deep integration with disciplines such as computer science, mathematics, cognitive science, neuroscience, and social sciences. With breakthroughs in technologies such as super-resolution optical imaging, optogenetic control, transparent brains, and somatic cell cloning, the development of brain and cognitive sciences has ushered in a new era, enabling large-scale and more detailed analysis of the foundational mechanisms of intelligence. AI will enter a biologically inspired intelligent stage, relying on discoveries in biology, brain science, life sciences, and psychology to convert mechanisms into computable models, while AI will also promote the development of traditional sciences such as brain science, cognitive science, life science, and even chemistry, physics, and astronomy.
The AI industry will flourish. With the further maturation of AI technology and the increasing investment from governments and industries, the cloud-based applications of AI will accelerate continuously, and the global AI industry scale is expected to enter a high growth period in the next decade. For example, in September 2016, consulting firm Accenture released a report indicating that the application of AI technology will inject new momentum into economic development, potentially increasing labor productivity by 40% based on existing foundations; by 2035, the annual economic growth rate of 12 developed countries, including the U.S., Japan, the U.K., Germany, and France, could double. A 2018 McKinsey report predicted that by 2030, about 70% of companies will adopt at least one form of AI, and the new economic scale added by AI could reach $13 trillion.
AI will promote humanity’s entry into an inclusive intelligent society. The “AI + X” innovation model will mature with the development of technology and industry, producing revolutionary impacts on productivity and industrial structure, and driving humanity into an inclusive intelligent society. In 2017, the International Data Corporation (IDC) pointed out in its white paper “Information Flow Leads to a New Era of AI” that AI will enhance operational efficiency across various industries in the next five years. China’s economic and social transformation and upgrading have a significant demand for AI, and driven by the demand in consumption scenarios and industry applications, it is necessary to break the perception bottlenecks, interaction bottlenecks, and decision-making bottlenecks of AI technology, promote the integration and enhancement of AI technology with various sectors of society, and establish several benchmark application scenarios to achieve low cost, high efficiency, and wide-ranging inclusive intelligent society.
International competition in the field of AI will become increasingly intense. Currently, the international race in the field of AI has begun and is becoming increasingly heated. In April 2018, the European Commission planned to invest $24 billion in AI from 2018 to 2020; in May 2018, the French president announced the “French AI Strategy” to welcome the new era of AI development and make France a strong AI nation; in June 2018, Japan’s “Future Investment Strategy 2018” focused on promoting the construction of the Internet of Things and the application of AI. Major military powers have also gradually formed a competitive situation centered on accelerating the development of intelligent weaponry. For example, the Trump administration’s first “National Defense Strategy” report seeks to maintain military advantage through technological innovation such as AI, ensuring that the U.S. can win future wars; Russia proposed in 2017 to embrace “intelligence” in military industries, enhancing the power of traditional weapons like missiles and drones.
The sociology of AI will be put on the agenda. To ensure the healthy and sustainable development of AI and to ensure that its development results benefit the public, it is necessary to systematically and comprehensively study the impact of AI on human society from a sociological perspective and to formulate comprehensive laws and regulations for AI to mitigate potential risks. In September 2017, the United Nations Interregional Crime and Justice Research Institute (UNICRI) decided to establish the first United Nations Center for AI and Robotics in The Hague to regulate AI development. The U.S. White House has repeatedly organized seminars and consultations on legal and regulatory issues in the field of AI. Industry giants such as Tesla have led the establishment of organizations like OpenAI, aiming to “promote and develop friendly AI in ways that benefit all humanity.”
Situation and Reflections
Currently, the overall situation of AI development in our country is good. However, we must also be clear that there are risks of overheating and bubbles in AI development, especially in foundational research, technical systems, application ecosystems, innovative talent, and legal norms, among other aspects, which still pose many noteworthy issues. Overall, the current state of AI development in our country can be summarized as “highly valued, promising situation, significant gaps, and optimistic prospects.”
Highly valued. The Party Central Committee and the State Council attach great importance to and strongly support the development of AI. General Secretary Xi Jinping has repeatedly emphasized the need to accelerate the development of a new generation of AI at events such as the 19th National Congress, the 2018 Academician Conference of the Two Academies, the National Cybersecurity and Informatization Work Conference, and the ninth collective study of the 19th Central Politburo. In July 2017, the State Council released the “New Generation AI Development Plan,” placing the new generation of AI at the national strategic level and outlining a roadmap for AI development in our country toward 2030, aiming to build an early advantage in AI and seize the strategic initiative of a new round of technological revolution. National ministries such as the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Science and Technology, and the Ministry of Education, as well as local governments in Beijing, Shanghai, Guangdong, Jiangsu, and Zhejiang, have introduced encouraging policies for the development of AI.
Promising situation. According to the “China AI Development Report 2018” released by Tsinghua University, our country has become the largest country in the world in terms of AI investment and financing scale, and our AI enterprises are at the international forefront in application fields such as facial recognition, speech recognition, security monitoring, smart speakers, and smart homes. According to statistics from the Elsevier literature database in 2017, the number of papers published in the field of AI in our country ranks first in the world. In recent years, universities such as the University of Chinese Academy of Sciences, Tsinghua University, and Peking University have established AI colleges, and the China AI Conference, which started in 2015, has successfully held four sessions with continuously expanding scale. Overall, innovation and entrepreneurship, as well as educational and research activities in the field of AI in our country are very active.
Significant gaps. Currently, our country is generally in a “catch-up” position in terms of innovative theoretical advancements in AI, with most innovations leaning toward technological applications. There are still significant gaps compared to world-leading levels in foundational research, original achievements, top talent, technical ecosystems, foundational platforms, and standards. Among the global top 700 AI talents, although China ranks second in the number of selected individuals, it is far below the U.S., which accounts for about half of the total. In 2018, market research consulting firm Compass Intelligence ranked over 100 AI computing chip companies worldwide, and no Chinese company made it into the top ten. Additionally, the AI open-source community and technology ecosystem layout in our country are relatively lagging, and there is a need to enhance the strength of technical platform construction and improve international influence. Our country’s enthusiasm and efforts in participating in the formulation of international AI standards are insufficient, and domestic standard formulation and implementation are also relatively lagging. There is a lack of in-depth analysis of the potential social impacts of AI in our country, and the process of formulating comprehensive laws and regulations related to AI needs to be accelerated.
Optimistic prospects. Our country has comprehensive advantages in market scale, application scenarios, data resources, human resources, smartphone penetration, capital investment, and national policy support for AI development. A report by global top management consulting firm Accenture in 2017 indicated that AI is expected to increase China’s labor productivity by 27% by 2035. The “New Generation AI Development Plan” released by our country proposed that by 2030, the core AI industry scale will exceed 1 trillion yuan, driving the related industry scale to exceed 10 trillion yuan. In the future development journey of our country, the “intelligent dividend” is expected to compensate for the shortfall of the demographic dividend.
Currently, it is a major historical opportunity for our country to strengthen AI layout, reap the benefits of AI, and lead the intelligent era. How to choose the right path for China, seize the opportunities for China, and showcase the wisdom of China amidst the booming development of AI requires deep reflection.
Establish a rational and pragmatic development concept. The development of any phenomenon cannot always remain at a high level; there are peaks and troughs, which is an objective law. Achieving autonomous intelligence and general intelligence for machines in any real environment still requires long-term theoretical and technical accumulation, and the penetration and integration of AI into traditional fields such as industry, transportation, and healthcare is a long-term process that is difficult to achieve overnight. Therefore, developing AI should fully consider the limitations of AI technology, recognize the long-term and arduous nature of AI reshaping traditional industries, rationally analyze the demands for AI development, rationally set development goals for AI, rationally choose development paths for AI, and pragmatically promote AI development measures. Only in this way can we ensure the healthy and sustainable development of AI.
Emphasize foundational research and original innovation. The cutting-edge foundational theories of AI are the cornerstone for breakthroughs in AI technology, industry innovation, and industrialization. Facing the critical point of development, to gain the final say, we must achieve significant breakthroughs in foundational theories and cutting-edge technologies of AI. We should follow General Secretary Xi Jinping’s call to support scientists in venturing into the “unexplored areas” of AI technology and strive for transformative and disruptive breakthroughs in the direction, theories, methods, tools, and systems of AI development, forming an original theoretical system of AI with international influence, providing leading theoretical support for building our own controllable AI technology innovation ecosystem.
Build a controllable innovation ecosystem. The AI open-source community and technological innovation ecosystem layout in our country are relatively lagging, and there is a need to strengthen the construction of technical platforms. We should be problem-oriented, focusing on key core technologies, accelerating the establishment of a new generation of AI key common technology system, and comprehensively enhancing the technological innovation capability of AI, ensuring that key core technologies of AI are firmly in our hands. We should pay close attention to preventing the “hollowing out” risk in the AI era, systematically layout, and focus on developing the “new core high base” in the field of AI: “new” refers to a new type of open innovation ecosystem, such as the integration of industry, academia, and research; “core” refers to core key technologies and devices, such as advanced machine learning technologies, robust pattern recognition technologies, low-power intelligent computing chips; “high” refers to high-end comprehensive application systems and platforms, such as machine learning software and hardware platforms, large data platforms; “base” refers to foundational theories and methods with significant original significance and technological driving force, such as brain-computer interfaces and brain-like intelligence. At the same time, we should emphasize the construction of AI technology standards and the testing of product performance and system safety. In particular, our country is at the forefront of the world in the application of AI technology, and we should master the discourse power in the formulation of international AI standards, and accelerate the process of AI-driven economic and social transformation and upgrading through the implementation of standards.
Promote a shared global governance. Currently, developed countries have controlled the upstream resources of the industrial chain through AI technology innovation, and the insurmountable technological gap and industrial barriers may further widen the gap in productivity development between developed and developing countries. Among developing countries, our country is expected to become a leader in global AI competition, and should layout and build an open, high-quality, affordable, and globally inclusive AI technology and application platform, coordinating with the “Belt and Road” initiative, allowing the “intelligent dividend” to promote the building of a community with a shared future for mankind.
Source: “Qiushi” 2019/04
Author: Tan Tieniu, Deputy Director of the Liaison Office of the Central People’s Government in the Hong Kong Special Administrative Region, Academician of the Chinese Academy of Sciences