On July 8, the State Council of the People’s Republic of China issued the Development Plan on a New Generation of Artificial Intelligence. This is a lengthy document, divided in six sections and several paragraphs. A translation produced by the fantastically efficient interns of the Foundation for Law and International Affairs is 27 pages in length. I have taken the time to go through the document in its Chinese and its English versions, to learn more about its contents, and to summarize it for my readers.
This is an extremely important document, outlining the blueprint for the development of China’s AI in the next thirteen years. As it is already clear to those in the know, the Plan has significant implications for the development of social credit, as well as the fields and the policies related to it.
Doomers and gloomers may object that documents as this one are programmatic, and that they should never be taken at face value, because the future is uncertain, and policy lines may change due to internal strife, crackdowns, policy struggles, factionalism, corruption, environmental crises, and so on.
Readers of this blog, however, may want to consider how vaguely similar arguments were made at the end of the 1970s, about China’s policy of reform and opening up to the outside world. Those old arguments, I guess, need not be commented upon. A picture can be worth more than a thousand words, and this is how China’s big cities can look, today:
In the following part of this post, I present a section by section summary on the State Council Development Plan on a New Generation of Artificial Intelligence. Your feedback and comments are welcome.
Inside China’s Plan on Artificial Intelligence
The first section of the Plan sets forth the opportunities and the challenges presented by AI.
The biggest opportunity lies in the development of AI itself, and in the fact the world’s economy is moving from an era of digitalization and from a ‘network economy’ to a model of economic growth and development based on automation. AI and automation already are the driving force of economic growth and development therefore, China should not miss this trend, and it should adopt necessary measures to promote the development of AI. All relevant measures should be promptly adopted in all relevant sectors of the economy, in order to facilitate the country’s economic development. The opportunities offered by AI are not limited to economic growth only: AI can find several applications in the most diverse fields of public administration, ranging from medical care to environmental protection, from construction to crime control, etc.
The development of AI doesn’t come without challenges. These are explored in pagraph 5 of Section 1 of the plan. “Artificial intelligence”, so it is written in the Plan, “can affect (…) global governance, which may lead to problems of changes in employment structure, impact law and social ethics, violate personal privacy and challenge international relations”. Out of consideration for these global challenges, the Plan makes a commitment to developing AI in ways which minimize global risks, and are safe, reliable, and controllable.
The Plan proceeds by summarizing and describing all the measures China has this far taken to develop AI, and then lists the sectors of the economy that await to be further developed, in order to ensure that the goals listed in the second part of the plan are met.
The second section of the Plan lists the ideological, political, and policy principles and the premises that will shape the development of AI in China. A deep understanding of this section is essential to know which path the development of AI will take, and which among the many possible different scenarios that could be planned, will more likely materialize. On the other hand, except for paragraph 3 this section is a long sequence of linguistic codes, policy shorthands, and various subtle references which, at a first sight, do not make sense to the average, untrained reader.
To know where AI in China will be heading, and to make a correct forecast of the likeliest scenarios, this section needs to be decoded using the right interpretive keys. Using interpretive keys that do not fit the lock of linguistic codes will lead the interpreter astray, with all the possible negative consequences a wrong understanding of this document may sort.
The third section of the Plan describes all the necessary steps China needs to make, together with its partners, in order to achieve the developmental goals listed under section II. China has to gauge current trends in the development of AI, to identify the best and most valuable opportunities, and to avoid the domestic and global challenges an unregulated development and use of AI could pose. To this end, it is essential for research on AI to take place in a cooperative and in an open way.
Paragraph A of the third section of the Plan is divided in four sections, which describe how scientific research on AI will take place.
First, research on AI will start by focussing on existing studies, and determining where the research frontier in AI lies. This stage of research will be mostly theoretical, and will focus on ways to bring the practice of AI in line with relevant theories. Once this stage of research will have been completed, China will focus on such fields as machine learning, and adaptive learning. A further step will be promoting interdisciplinary research, to study the possible applications of AI to neurosciences, to quantum physics, psychology, mathematics, sociology, and other disciplines. Legal research will be part of these interdisciplinary efforts.
Second, a later stage of research will focus on improving human-computer interaction through the use of means such as algorithms, and hardware components. Human-computer interaction is an essential component of day-to-day activities as performing internet searches, knowledge representation, data analysis, data mining, the use of social media, but also using internet-based and mobile applications. Other applications of human-computer interaction involve GPS technology, self-driving cars and railway carriages, UAVs, industrial automation processes, etc. Next come the sectors of virtual reality, and natural language processing technologies, to ensure a better communication beetween man and machine.
Third, China will construct a platform on artificial intelligence. While it is not yet known what form the platform will eventually take, its goal will be integrating various applications AI may take across different sectors, and promoting the innovation of several industrial sectors, included scientific research.
Fourth, to make this vision become a reality, China will heavily invest in human resources, in R&D, and in education. Such an investment strategy will have a global dimension, in that it will involve attracting the best and brightest minds to conduct research and educational activities in each one of the fields related to AI. To this end, China will introduce a set of policies and conditions favorable to foreign experts, and researchers. Universities, research institutes, and enterprises will be encouraged to focus on all those disciplines related to the development of AI.
This section of the Plan contains three boxes where a more detailed description of research fields and topics is provided.
Paragaph B of the third section of the Plan states the goals to use AI to promote innovation, and to construct an economic system where AI will play a major role. This paragraph is divided in three sections, which explain how these goals will be reached.
First, it will be necessary to develop existing forms of AI, and to create new forms of AI through the construction of internationally competitive industrial clusters. These industrial clusters will be: the hardware and software industry, robotics, the vehicle industry, through the production of intelligent self-driving vehicles, virtual reality, smartphones, and smart gadgets such as watches, headphones, glasses, and the internet of things.
Second, AI will have to be integrated with industrial innovation. This will mean applying AI to the sectors of manufacturing, agriculture, logistics, finance, business intelligence, construction, household appliances, etc.
Third, enterprises innovation will have to be promoted. This will involve encouraging enterprises to apply AI and related technologies to production, management, logistics, marketing, as well as productive activities. Enterprises should use big data technologies to facilitate processes of production, distribution, and management. A significant role in industrial innovation will be played by voice and facial recognition technologies, robotics, and virtual reality. Therefore, enterprises should be encouraged in innovating, and in participating in international research projects.
Fourth, to facilitate innovation China will introduce pilot projects in th most promising areas, it will construct national industrial parks devoted to AI, promote the creation of innovation clusters, and create national AI bases, where academics and professionals will cooperate in R&D activities.
Paragraph C of the third section of the plan examines the applications of AI to the societal field. The stated goals of this paragraph are improving citizens’quality of live, using AI to perform repetitive, tedious or dangerous jobs, to employ human resources in those industries and tasks that require greater analytical capabilities, creativity etc. The application of AI to the societal field will involve the following four fields of society:
First, the education, health care, and social security systems. AI will find applications in teaching, the management of HR employed in the education industry, and in the creation of online learning platforms. AI will also be used to in the health care sectors, to perform surgery, to create wearable and biologically compatible monitoring devices, and in such fields as image recognition, the recognition of pathologies, the prevention of epidemics, etc. Big data and the internet of things will find applications in the pension system, as well as in the development of audio-visual aids, prostethics etc.
Second, the social governance field. Here a key role will be played in administrative management, the judiciary, urban management, and environmental protection.
Third, the field of public safety. AI will be used for information analysis, personal identification, biometric identification, as well as in criminal investigation. The field of public safety is however not limited to these activities, and includes also the use of AI in the monitoring and response to natural disasters, such as earthquakes, floods, etc.
Fourth, AI will find applications also in the broader field of ‘integrity’ and ‘trust’. This field obviously relates to the use of AI in all those areas and activities related to the construction of a social credit system.
Paragraph D of the third section of the plan examines the use of AI in civil-military integration. This paragraph outlines all the measures relevant to reach the goal of integration between the dual uses of AI, and it contains a box where all the measures and steps needed to construct an ‘intelligent infrastructure’ are listed.
The fourth section of the Plan makes provisions about the allocation of all the financial resources necessary to guarantee implementation of the measures listed under the first, second, and third sections. It is important to notice how the development of AI will rely on global resources, understood both as domestic and international investments, and other incentives. This section is divided in three paragraphs.
The first paragraph outlines the mechanisms that will be used to support the development of AI. These will involve both the government, and private enterprises. China will establish a ‘market-oriented’ fund to ensure the development of AI, and use venture capital, capital funds, and capital market financing to fund domestic and international research projects.
The second paragraph describes how China will construct leading innovation centres such as enterprise laboratories, engineering laboratories, and so on. This paragraph specifies how private enterprises will be a major driving force behind all these efforts.
The third paragraph is devoted to the global coordination of efforts to further develop AI. China will encourage domestic enterprises to ‘go out’. But, international enterprises will be encouraged to set up R&D activities in China too. The development of AI, so this section says, will be weaved into the One Belt One Road, and it will involve the creation of international organizations and standards on AI.
The fifth section of the Plan lists all the measures needed to regulate the AI sector. This section is divided in six paragraphs.
First, China will develop a regulatory framework to promote the development of AI. This regulatory framework will include national legislation, but also administrative laws, regulations, and ethical codes to be adopted by all the enterprises active in the AI field. Further, China will develop accountability systems.
Second, enterprises choosing to invest in the AI sector will benefit from preferential tax policies, and fiscal incentives, as well as from a more stringent standard of IPRs protection.
Third, safety and monitoring systems will be developed, to prevent and manage all the risks an unregolated, uncontrolled development of AI may pose to employment, to ethics, accountability, to the domestic and global economic system, etc.
Fourth, fifth and sixth, a greater awareness of AI will be raised in society, by encouraging universities, research institutions, and private compenies to invest greater resources in the creation of platform needed to educate the public in basic notions about AI, and to train a generation of competent and internationally competitive scientists, programmers, researchers, etc.
The Plan closes with a short, sixth section describing the division of responsibilities among various organs of the Party and of the State, and providing terse information about the need to start pilot projects in the development of AI.