AI in China: Technology for development?


At various levels the nature of the focus given and consequently the developmental contours of AI as a technology in China are very new, as they arise from a new set of opportunities and challenges. But there are also parallels with past experiences of regulating on and promoting the growth of such ‘new’ or ‘emerging’ technologies. This essay seeks to understand how the view of the Chinese state and the consequent policy directions it pursued in the case of AI differed from those of other emerging technologies, by training focus on the biotechnology sector from the 80s up until the 2000s, with the theoretical frame of the state as a key actor in the technology-for-development paradigm.

In many ways the story of the newfound attention that AI is receiving begins with a rather dramatic event in March 2016 — the defeat of Lee Sedol by Google’s DeepMind in a game of Chinese origin, AlphaGo (Ding, 2018, p-7). Following this, AI became a hot button national issue, with not just the citizens or the state but even the PLA waking up to its deeper potentials.

This is not to say that AI had not been of importance before. Many of China’s provinces, and private organizations had been investing in their AI capacities for quite a while — the city of Tianjin for example had allocated 5 billion USD to the development of the AI industry a month before the State Council’s plan of 2017(Ding, 2018, p-8). But this single event produced reactions from nearly all quarters — investing in and leading in AI became a question of national dignity and pride. One of the most interesting statements was from the PLA:

“Per testimony before the U.S.-China Economic and Security Review Commission by Elsa Kania, the PLA anticipates the advent of artificial intelligence will fundamentally alter the character of warfare, ultimately resulting in a transformation from today’s ‘informatized’ ways of warfare to future ‘intelligentized’ warfare.” (Ding, 2018, p-13)

It would be worth thinking through the differences implicit and explicit, in the contrast here between ‘informatized’ and ‘intelligentized’. Informationization is a term that has had wide use in Chinese policy writing around technology. Jiang Zemin’s endorsement of the term was quite strong when writing about the merits of developing the IT sector to produce an informatized economy and society, detailed in the previous essay. But what does it really connotate, particularly in this context, where it is contrasted with intelligentized or intelligentization?

Information is a valuable tool in the larger development process, but is understood as ultimately needing ‘processing’; i.e., it needs to be made sense of and acted upon, both of which are activities traditionally involving significant human intervention, or ‘intelligence’. This is unlike an ‘intelligentized’ set of devices, procedures, etc. where the human element can be minimised and preferably removed completely — the machine can beat a human at a game of Go and possibly then, on the battlefield, with equal ease, with no human to hold its proverbial hand. Naturally, this halo around AI’s perceived potentials shapes the way several interest and power groups such as the larger scientific community, the bureaucracy, the military and even the general public see it. This has resulted in a shift not just from the command-and-control, domestic market centric or half-baked liberal approaches of the previous decades, as seen in the larger IT sector of the 80s and 90s, and agricultural technology before that, but rather a policymaking trajectory which shares characteristics with each of them, and is still unique.


While AI is a lazy blanket term used to speak of many things, Ding has tried to define AI through its components as “(1) hardware in the form of chips for training and executing AI algorithms, (2) data as an input for AI algorithms, (3) research and algorithm development, and (4) the commercial AI ecosystem” (Ding, 2018, p-4).

In China’s case he admits, except for easy access to humongous data reserves, nearly all three other components were lacking. So, China’s AI Potential Index, a measure of a country’s “overall AI capabilities” is a measly 17 compared to the 33 of the US, which invests in and possesses huge reserves of research and talent and is still the leader in the present AI commercial ecosystem (Ding, 2018, p-5). In a section ahead Ding details how this was not just an academic assessment, the Chinese state was also keenly aware of its own modest position in the AI scheme of things, compared to the US, the UK and even “Robot Superpower Japan” and “Ambitious EU” (Ding, 2018, pp 11–12). Another facet of the AI industry in China is that its most important areas of research investment and activity are in NLP (31%), Voice (31%) and Facial Recognition(25%) and Computer Vision (Lewis, 2016, p-3). As Lewis notes “ Facial recognition makes up 35% of all AI applications in China and it is in this area that some of China’s most well-known, and globally controversial, AI unicorns, such as Sensetime, Megvi Face, and Yitu, have emerged” (Lewis, 2016, p-3).

In a nutshell, China was far behind the leading AI powers of the world, at least as of 2017.

But that did not prevent envisioning a future that was radically different — the State Council AI Industry Plan (2017) document states that “By 2030, China seeks to become the world’s “primary” AI innovation centre, with a core AI industry gross output exceeding

RMB 1 trillion (USD 150.8 billion) and AI-related gross output exceeding RMB 10 trillion (USD 1.5 trillion)” (Ding, 2018, p-13). This was the culmination of a three-stage process which begins with bringing China’s AI capacities “in line” with global leaders by 2020, and then by attaining a “world leading level” in at least some core industries by 2025 (Ding, 2018, p-10).

While the 2017 Plan was a national level plan, that was modelled on similar plans by the US and other countries, it was not the first such national plan, and was the result of major bureaucratic and political deliberation and attention over a period of time. The 13th five year plan (2016–2020) “highlighted development of AI as 6th among 69 major tasks for the central government to pursue”, and similarly the three year AI implementation plan 2016–2018 was a pivotal moment where AI was being linked to the ambitious “Internet Plus” program, cementing its important position in turning China into a “digital powerhouse” (Ding, 2018, p-9). There is also Robotics Industry Development Plan (2016–2020) which is significant in that it laid out how China was to reach the goal of becoming a world leader in robotics, producing an ambitious 10,000 of them annually by 2020 (Ding, 2018, p-9). Besides these grand national plans, the phases of AI development have also been influenced by a degree of decentralisation — different regions and cities, like Tianjin mentioned before, were able to set up funds and position themselves as clusters of AI innovation, development and manufacturing. In fact the emphasis on robotics and particularly their manufacturing, point to an ability to perceptively develop a strategy based on an understanding of one’s own capacities and strengths.

Another facet of AI much like IT discussed in the previous essay is the previously mentioned halo around its potentials, which is revealed in the dilemma over what exactly constitutes ‘core AI’ industries and entities and those that are only ‘AI-related’ or ‘AI-enabled’. There are several points at which this difference begins to blur, what Ding calls slipperiness, arising from the “omni-use potential” of AI (Ding, 2018, p-11). On one perhaps more clear end, it is possible to refer to industries like the manufacture and design of chips and robots, deep learning and so on as ‘core’ or what can be referred to as “industry-agnostic” or “general purpose” AI which finds application across a variety of sectors. On the other end smart cities, robots and wearables would be labelled as ‘enabled or related’. This is understood by and incorporated in the “Internet Plus” AI Three Year Implementation Plan discussed above, by separating nine “core” AI technologies from 8 related or non-core technologies (Ding, 2018, p-9). So there is evidence of the steady incorporation of AI in China into the larger scheme of economic development and growth. In fact to Ding, “economic benefit is the primary, immediate driving force behind China’s development of AI” (Ding, 2018, p-6). This is in no small part because globally, China would benefit the most in GDP terms by incorporating and developing AI enabled economic and industrial systems, doing so would logically compound its already enviable growth rate and trajectory. Besides contributing the development of such allied or enabled sectors, the core AI industry was itself contributing a significant 15 billion RMB in 2017, which was sought to rise nearly tenfold, to 150 billion RMB in around 3 years by 2020, according to the state’s ambitious estimates (Ding, 2018, p-13). As Ding notes, this was not an unwarranted figure, with the global industry being predicted to rise to USD 126 billion in the same year from from USD 644 million in 2016, which explains the high Chinese estimate (RMB 150 billion is roughly 60.3 billion USD) (Ding, 2018, p-10).


The prioritisation and integration of new technologies as tools of extreme significance in economic growth is one of the many consistencies in the way China approached the development of AI. This can be better illustrated by exploring the parallels between the contours of other industries such as the biotechnology and IT sectors with AI.

One of the first key characteristics would be ambition. In the case of the IT industry, this was visibilised by no less than the premier, Jiang Zemin, an electrical engineer with a deep and abiding passion for the ‘informatisation’ of China, which he saw as key to its ascent in the global order, discussed in the previous essay. In the biotechnology sector where China sought to possess the largest capacity of plant biotechnology outside the America (Huang and Wang, 2002, p-122), this ability to look at a sector as not just for national ‘development’ but also a premium seat at the global economic table, is one of the unique aspects of Chinese technology development. It could be rather crudely or dramatically characterised as a ‘hunger for power’ or leadership, global dominance. Like IT, biotechnology received significant attention in the 80s, akin to what AI is receiving now. The state was able to devolute massive funding and take swift decisions — for example the approval of field trials and commercialisation for no less than 12 crops across cotton tomato and petunia, almost immediately after the National GMO Biosafety Committee was set up in 1997! (Huang and Wang, 2002, p-122). For nearly a decade and a half, this hunger would propel China to become the world leader it had set out to be, when quite suddenly the media and the industry itself within China would announce the “winter of biotechnology” (Huang and Wang, 2002, p-122).


This was characterised by a near reversal in the previously fast paced approval system portending a general distrust of Genetic Modification, with several regulatory bodies on biosafety and certifications coming up seeming to throttle advancement in the field. However, as Huang and Wang note, this closing of the chapter of biotechnology was premature and not entirely true. Like in the case of AI, biotechnology and more generally development in agrarian technology would come to be associated with improving human life. They quote Jiang Zemin in 2000, as saying:

“We are also very much concerned about these…. (biosafety and the harms of GM technology).I think it is important to uphold the principle of freedom of science. But advances in science must serve, not harm humankind. The Chinese government is now mulling over new rules and regulations to guide, promote,regulate, and guarantee a healthy development of science. I believe biotechnology — especially gene research — will bring good to humanity.” (Rubenstein, 2000 in Huang and Wang, 2000, p-123)

What is another interesting parallel to note is that of biosafety in the case of biotechnology and the ethics and privacy debate around AI. In both instances as Ding and this paper also find, there has been at times a simplistic mischaracterisation of the Chinese position. When the Chinese initially prioritised biosafety regulations and streamlining in the early 2000s, that was at once attacked by the industry for being too restrictive and producing the aforementioned ‘winter’. However global bodies such as Greenpeace and other environmental bodies would find that they were validated, after years of warnings about the hazards and risks of the large scale and relatively quick materialisation of China’s GM aims. Similarly there is much opprobrium over China’s moves in the AI arms race, and its apparent lack of substantive discussions around safety and ethics in AI. To Ding, some of this is genuinely unfounded. His assessment of China’s AI capacities mentioned before, can be simplistically understood as it being half as ‘powerful’ as the US. Furthermore, in the case of ethics, as with biosafety, there have been increasingly more discussions around the safety and capacities of AI. Again, like in the case of biosafety where regulatory and bureaucratic bodies were set up to research and regulate on the future directions and consequences of GM research and applications, (Huang and Wang, 2002, p-127) ethics and safety in AI has also received significant state attention. In the very State Council Plan of 2017 which laid out a concrete timeline of the growth of the AI industry, there was also a timeline laid out for establishing and putting into force regulatory measures and safety standards. It outlined that “by 2025, China will have initially established AI laws and regulations, ethical norms, and beginnings of AI security assessment and control capabilities; and by 2030, China will have constructed more comprehensive AI laws and regulations, as well as ethical norms”(Ding, 2018, p-30). Tencent published a book on the same where Ding notes that there are “chapters that are relatively proactive in calling for AI safety mechanisms”(Ding, 2018, p-30). However, one of the limitations of this emerging debate around the unique risks of AI technologies and their misuse is that much of it is still driven by Chinese institutions within China, i.e. there seems to be a lack of engagement with the global discourse around AI safety and ethics(Ding, 2018, p-30). A notable instance of this changing for the better is the involvement of a number of AI Chinese researchers in producing the translation of the IEEE’s Ethically Aligned Design Report(Ding, 2018, p-30). There are also as Ding further elaborates a diversity of viewpoints and strands of thinking around how to govern AI development. He takes, on the one hand, the example of Dean Shenn who stated that “AI development was an immutable social trend that should be embraced rather than excessively worried over”(Ding, 2018, p-31). On the other hand, Guobin Li, a prominent intellectual and the then President of the Beijing Research Institute for Communication Law, who passionately advocated the building of legal and policy capacities in understanding and regulating on AI(Ding, 2018, p-31). The ethics discourse is particularly important because of its increasing dominance in the context of international standard setting. So, the development of AI and its economic capacities was to go hand in hand with the construction and promotion of a framework or code of AI ethics. This was acknowledged as not just for the purposes of achieving a techno-utopia but also “seize the strategic highground” in the global AI order (Lewis, 2016, p-1).

“Data, Data, Data”:

Like ethics, a debate is also unfolding over data privacy protections. Nearly all stakeholders — the general public, big tech companies and the bureaucracy and government at various levels, find themselves along a spectrum of opinion on the subject. Researchers from Tencent are able to clearly attribute the shiny success of American tech to the liberalised yet safe environment there, that rides on the back of copyright laws, which goes in hand with the strength of institutions and innovation building. Thus they note that:

“If there is no government data liberalization policy, many AI applications will become ‘water without a source, a tree without roots.’ It can be said that the issue of data liberalization is a pain point in the development of AI in China and needs to be elaborated upon in a more comprehensive and in-depth manner in the strategy” ((Ding, 2018, p-31)

At the same time the nuanced Personal Information Security Specification that China released in 2018 was being heralded as even more comprehensive and powerful than the GDPR, which was at the moment the global standard! (Ding, 2018, p-31) Again like in the case of ethics, Chinese firms and also the state have come to see setting standards in privacy as equally important, to improving their position in the AI ladder, as in developing sheer capacity. This is of note, because historically, first entrants into any field, normally do not get to set standards, and even if they do, it is not always a priority for them. By engaging with the questions of ethics and privacy, China is very deftly and definitely playing the long game, and is reaffirming its very clear strive for techno-independence.

Talent transfers:

Besides ethics and safety concerns there are also similarities in the development of talent and human resource in biotechnology and AI. In both fields concerns were expressed over the lack of this critical resource. In the case of AI concerns of how a practical lack of all resources besides massive data troves, meant that China couldn’t really hope to become a leader in the field in the recent past, have been discussed in the previous sections. Similarly in biotechnology, China could depend on a steady availability of massive land, labour and a market for crops with enhanced capacities. However in both fields the development of human talent and indigenous research was critical to their growth. So, emphasis was placed on a dual pronged strategy of both boosting global flows of talent, to and from research institutes and corporates in the West and China. These were in part borne from the creation and return of Deng’s proverbial “turtles” — the Chinese students and academic community who were trained and educated in the institutions of the US and other Western countries, and also the promotion and materialisation of the many opportunities that awaited them back home . In the case of biotechnology, like with AI, there had been sustained public investment in generating modern globally competitive research and research environments right from the early 80s. As of 2000, there were nearly 5000 researchers across nearly 50 institutes working in both plant and animal biotechnology (Huang and Wang, 2002, p-125). In the period that Huang and Wang studied, i.e, from the 80s to the mid 90s and early 2000s they noted that there had been a tripling in the number of researchers and a tenfold increase in those who held a PhD in the field, signalling that the rise in numbers was to continue(Ding, 2018, p-127). Proof of this, at the time of Ding’s writing, was in the advanced capacities of cloning techniques and even research on viral epidemics, that China had come to possess(Ding, 2018, p-15).

Another aspect to note beyond the attitudes, ambitions and that governed AI and biotechnology development is the public-private investment and interaction ecosystem.


The most important organization which governs much of the policy making around technology in China is the Ministry of Science and Technology. In fact, the beginning of a sustained policy direction and prioritisation for AI was laid in the “National Medium- and Long-Term Plan (MLP) for the Development of Science and Technology (2006–2020).” This plan laid the groundwork for the Chinese government’s involvement in AI development, seen in the allocation of nearly 500 Billion RMB (75 Billion USD) to the development of “vanguard science and technology” of which core AI and AI-supported projects were key components (Ding, 2018, p-14). Even the “Made in China 2025” initiative and the “Artificial Intelligence 2.0” designation follow from the principles of indegenised innovation and sustained growth of advanced manufacturing capacities, found in the MLP of 2006 (Ding, 2018, p-14). In the case of biotechnology, the state would prioritize it as one of the seven critical recipients of funding from the prominent 863 program, designating it as a similar national priority(Huang and Wang, 2002, p-126).
So one of the contentious periods in this bureaucratisation came from the schisms emerging between the several groups involved in the development of the AI MLP. The Chinese scientific community and the bureaucrats of MoST at the time began to defer in their choice of the mega projects to be prioritized. This was in small part because of the fact that a massive 2000 people were involved in this drafting process (Ding, 2018, p-14). In the case of AI policy too, the various plans give different limits of powers and degrees of responsibility in governing and driving the growth of AI. The “Internet Plus and AI Three Year Implementation Plan” targeted the NDRC (NAtional Development and Reform Commision), the MoST, the MIIT (Ministry of Industry and Information Technology) and the Cyberspace Administration of China in devolving the responsibilities of AI development, whereas the 2017 State Council Plan gave authority singularly to MoST under which a new AI Plan Implementation Office was to come up and function (Ding, 2018, p 14–15). This infighting as Ding states is unlikely to reduce in the future, with the number of agencies involved rising to 15 (Ding, 2018, p-15). It is clear that multiple bureaucratic organizations, companies and other stakeholders are battling for prominence and their share of the AI pie.

This bureaucratization is a key characteristic of the way the Chinese go about pursuing development in a particular technological sector. In biotechnology too, the Ministry of Agriculture, the Chinese Academy of Sciences, The Education Ministry and the Forestry Bureau would come to be the main bureaucratic authorities governing the sheer number of staff, laboratories, funding capacities, institutional and organisational tie ups with universities, etc. that they possessed was enormous.

(Huang and Wang, 2002, p-126)

As Huang and Wang note, the growth of this bureaucratic apparatus signalled the opposite of ‘winter’ in biotechnology, with the state presence in the sector being built on the back of, and also paving the way for heavy investments and capacity building.


Speaking of investment and activity in the sector, while the efforts of the public sector has been robust, the private sector was not to be left behind in the case of AI, unlike in the case of biotechnology where investment was predominantly state driven. The former has been delineated to an extent in the above sections. But for the private sector in AI the BAT trinity — Baidu, Alibaba and Tencent — are the prime actors, along with Huawei, another tech giant. While each of these constitute the Chinese big tech and are not strictly AI only companies — Baidu is best known as the Chinese Google equivalent, Alibaba as an ecommerce giant rivalling Amazon, Tencent as messaging platform, and Huawei as a chip manufacturer — each of these is investing heavily in their AI capacities, particularly in their areas of advantage and operation to propel their continued competitive growth — “Alibaba in retail, finance, and entertainment marketing; Baidu in search and AI applications, especially in autonomous vehicles, Tencent in education and social, and Huawei in hardware through its phones and AI chips” (Lewis, 2016, p-3). As Lewis notes, the reality of this rather uneven landscape is that nearly 65% of 190 Chinese companies are recipients of investments for these four companies (Lewis, 2016, p-3). The public sector, besides the various plans and budgets outlined above, is able to significantly devolve funding through the Government Guidance Funds (GGF)(Lewis, 2016, p-4). The decentralised nature of AI development, where nearly half of all China’s provinces have their own AI development plans and budgets, shows that there is a strong degree of federalisation even if the broad policy direction and aims are set very firmly by the central bodies like the MoST. (Lewis, 2016, p-4) These GGFs were born out of an experiment of Beijing Municipal Government in 2002 and subsequently recognized by NDRC in 2008 number around 800–1000 across the country and have a fundraising total of around 5 trillion RMB (Lewis, 2016, p-4). Besides investing in R&D to boost innovation in advanced manufacturing, chipsets, big data, etc. these GGFs invest in companies by taking an equity share (Lewis, 2016, p-4). They form part of the effort to boost innovation and funding in areas of high risk and entry barriers, like the example that Lewis cites of the semiconductor industry, which are essential to larger growth capacity in AI and allied technologies.

‘TECHNOLOGY FOR DEVELOPMENT’ — Techno-Independence/Utilitarianism:

Lewis, like Ding also saw the AlphaGo defeat of Lee Sedol in 2016 as having “lit the ignition of the Chinese combustion engine that has since stayed in 6th gear, driving an ambition to first catch up to and then surpass all others as the world’s leading AI power” (Lewis, 2016, p-1). It is rather remarkable to see the extent to which an event like this should drive national policy on technology development. In many ways, this is part of a pattern which forms the evidence of the top down centralised rigidity that characterises Chinese decision making around technology. Further as Lewis and this paper also documents the historical and geopolitical ambitions circumscribe the development strategies of AI (Lewis, 2016, p-1). The Chinese state believes that true techno-independence can only be achieved by not just indigenising technological capacities and manufacture but also leading and dominating the world in that given technology. This again, is a consistent strain of thinking which is seen right from the days of Deng, when China survived on shamefully low quality imports from other better endowed countries (the “ghetto speakers” incident discussed in the previous essay) — developing indigenous capacities in advanced technology, was not just an economic necessity in a global competitive world order, but was necessary for something far more primal — dignity. To this effect the State Council Plan of 2017 was keenly aware of the unique challenges and advantages that faced the development of world class AI capacity in China. As Ding had elaborated, save for massive troves of data, China possessed none of the other key ingredients. The 2017 Plan itself mentioned as Lewis noted that “ever since the first industrial revolution it has consistently played catch-up to the West, particularly the US, lagging behind in patents, talent, and scientific research” (Lewis, 2016, p-1). So, China massively amped up its efforts in generating original research capacities, becoming the preferred destination for world class AI talent and a thriving public-private industrial ecosystem, with primacy in standard setting. That last aim was centered not just around technical specificities but even ethics and data privacy, discourses which were still maturing in the rest of the world. These efforts have paid off in some ways in that the Chinese private sector has come to be seen as globally pioneering in AI technology across a multitude of sectors from in diverse areas, as discussed in the previous section, and also in the increasing quantity and noteworthiness of Chinese research efforts in AI — nearly a quarter of all papers published on China are of Chinese origin(Lewis, 2016, p-2).

However, there are limitations to such top down strategies. To begin with, we see that the private sector has very narrow priorities. This is reflected in the fact that facial recognition technologies account for the biggest share of core AI industrial activity as discussed in the initial sections. Additionally, even the federalism, spoken of earlier, is problematic as government bodies compete with one another at the province and city levels to be seen as digitised and better at managing their areas of governance. This means they are more than willing to spend on AI technology, without deliberating too much on the kinds or specifics of the technology itself — the much maligned investments in CCTVisation too follows from this. Since most firms sell Facial Recognition tech, and that is AI, having CCTVs with advanced Facial Recognition capacities becomes seen as an improvement, as progress — technology for development. At these points it is valid to ask, “whose development is it really”? Is it the fragile ego of the Chinese bureaucracy which prompted the AI-sation of development, or is it really the people, who wish to see China as a “digital powerhouse”? What worth could such shiny technology have in the lives of millions of China’s poor? When the State Council Plan acknowledges that the state must be prepared to deal with the “social aftershocks” of an even more accelerated job losses and wider “digital divides” in the aftermath of such AI-sation, who is it speaking to? Who is it who will face that “social aftershock” (Ding, 2018, p-33)? What will “being prepared” for that mean? The very same document goes on to acknowledge that “AI will also play an “irreplaceable” (不可替代) role in maintaining social stability” (Ding, 2018, p-33). While Ding talks of the “Digital Leninism” that some Chinese thinkers this will produce, where AI will form a key part of the predictive policing effort (Deng, 2018, p-34). On the other hand there is the very real and to some extent already real possibility of AI and AI enabled technology, like the CCTV/Facial Recognition complex discussed above, shrinking the already almost non-existent space for resistance and opposition from the lowest denominators of the population.

This is just the surface of questions that can be asked, to unpack the larger risks of AI in the sheer scale that China is pursuing and seeks to pursue would be beyond the limits of this paper.

Beyond just these aspects, Lewis also goes on to critique the larger pattern of investment and the framing of regulatory interventions and mechanisms that is seen in the case of both Biotechnology as Huang and Wang illustrated, and in AI as Ding illustrated — that of “throwing money” (Lewis, 2016, p-3). Like in biotechnology, AI too has been showered with all sorts of massive funds, with Ding drawing parallels between the 2017 Plan and the 863 program of the 80s and 90s. Lewis quotes Roger Creemers, where he basically thinks of the MLP as a short-sighted and naive “Santa’s list of desiderata and objectives, but with little insight into how these should be achieved other than by throwing money at the problem” (Lewis, 2016, p-3). Further, he quotes Matt Sheehan from Marco Polo (Lewis, 2016, p-4) as saying:

“The hope is that if local officials cough up a sufficient number of these gifts — factories adopting smart robots, new research centers pursuing natural language processing, autonomous agricultural drone demonstration projects — they will eventually add up to the plan’s headline goal: global leadership in AI.”

The ambitious and well endowed GGFs also bring about cause for concern — there is as yet no assessment of how well they have actually performed in terms of propelling innovation or in turning high-risk investment areas into attractive ones. So it would be equally valid to be concerned that they are simply demolishing the scope for private sector investments and are essentially an ineffective use of state capital. Lewis notes that such an approach could in fact have the undesired effect of proving wasteful and actually doing nothing other than reducing the scope for a competitive market.

Besides investments the significant bureaucratic apparatus that gets developed is also contentious, in that it is difficult to ascertain if it is having the desired effect. As in the case of the drafting of the 2006 MLP, where nearly 2000 heads came together, and in the recent devolution of powers to various agencies, discussed in the section on bureaucratisation, infighting and competition between these bodies is a real pain point. What is disturbing is also that the priorities of this ‘technology for development’, gets decided by a very real federal version of “peer pressure” and what will win the most appreciation in bureaucratic eyes, not what may objectively or factually be the best for the sector or the industry itself.

Similarly there is the question of protectionism and the inability of Chinese companies to compete globally. Like in the case of software companies discussed in the previous essay, the Chinese totalitarian mentality, to govern as much as possible, goes against the direction of growth for indigenous companies. In the case of data, like Tencent researchers has noted, a global exchange is necessary for truly matching up to Western big tech. While it is noteworthy that the discussion on privacy and protectionism includes significant participation and acknowledgement of diverse opinions from all stakeholders — the state, big tech and the general public — it is still an ongoing process.


To conclude, while AI shares many characteristics in its initial trajectories with that of other industries such as biotechnology and even the software industries, it also has several unique features. On the one hand the state retains the ambitious top-down model of aiming for global domination, setting aside massive funds and building immense bureaucratic apparatus, while on the other, the participation and power of the private sector is very vital here, and drives in large part the priorities and capacities of AI being developed. While China is still a modest player, having only around half the American AI potential, its ability to produce rapid growth is testament to how it ramped up the talent and investment in original biotechnological research and is most forcefully doing the same in AI. Its pattern of development can be critiqued firstly on the grounds that it is not entirely clear what it seeks to pursue in the name of AI development, or that it fully understands the complex risks and challenges of doing so. Secondly, it doesn’t really see through the faults and loopholes in its own strategy beyond the ability to devolve large sums of money and power to subnational bodies. And lastly on the questions of enhancing data and talent flows, two vital components of AI growth and development, it is still to concretise its positions.

What is noteworthy is that it is not waiting to become a leader, or even to get a foot in the door, to account for questions of ethics, privacy and standards settings. It is playing to its strengths by making AI a provincial and even city level priority, not just some vague national fancy, and also focussing on manufacturing capacity among other already existing strengths.
China therefore remains one of the key ascending challengers to the Western technoeconomic hegemony. So while mixed and rarely one-tone, the trajectory of AI development in China, which has only just begun, is one to keep watching intently.


Ding, J. (2018, March). Deciphering China’s AI Dream: The context, components, capabilities, and consequences of China’s strategy to lead the world in AI. Center for the governance of AI.

Huang, J., & Wang, Q. (2002). Agricultural Biotechnology Development and Policy in China. AgBioForum, 5(4), 122–135.

Lewis, D. (2016, March 16). China’s Techno-Utilatarian Experiments with Artificial Intelligence. Digital Asia.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store