With AI poised to be the next disruption of business and the digital environment, China has set its sights on being the leader in the field. The effects of automation and AI, alongside machine learning and other digital enhancements like Big Data have already been well discussed, from enhancing the work we currently conduct to directly replacing us.
AI investment is rapidly growing, McKinsey and Company estimate, in their latest report on AI, estimate tech giants, including Google and Baidu, spent between $20bn to $30bn on AI in 2016, with 90 percent of this spent on R&D and deployment and the remainder 10 percent on AI acquisitions. Moreover, Private investors are also contributing to the ecosystem, with estimates that venture capitalists invested $4bn to $5bn in AI and private equity firms invested $1bn to $3bn in 2016, three times as much as in 2013. Further, on company usage in general, the survey found that ‘early AI adopters that combine strong digital capability with proactive strategies have higher profit margins and expect the performance gap with other firms to widen in the future.’
Yet, ‘corporate M&A is the fastest-growing external source of funding for AI companies,’ with large giants competing to secure the specialised talent that exists. The act of ‘acqui-hiring,’ is common in Facebook, Alibaba and Google, as firms seek to lead AI innovation, will more than 100 deals being concluded since 2010, 24 of which are Google’s alone.
China’s New Vision
July 20th 2017 saw China’s State Council issue the ‘Next Generation Artificial Intelligence Development Plan,’ outlining an ambitious and determined objective, for China to lead the world in AI, the premier global AI innovation centre by 2030. The agenda has a three pronged approach, tackling key problems in research and development, pursuing a range of products and applications, and cultivating an AI industry, with each prong receiving a deadline. Accenture Plc estimated that AI could increase China’s annual growth rate by 1.6% point to 7.9% by 2035, adding over $7trn in GDP.
2020 is purported to see China’s overall tech progress match the pace of the global effort, creating an environment conducive for the next generation of AI development. At this point, the target is to exceed $22bn in value, with AI related fields collectively valued at over $145bn. 2025 sees China progress further, ‘with AI becoming a primary driver for China’s industrial advances and economic transformation.’ Lastly, the 2030 vision sees China as a world leader in AI development, concluding the transformation with an AI industry valued at $148bn, and related fields collectively valuing at $1.48trn.
The plan is predicated on major breakthroughs occurring within the first stage to spearhead innovation in the following stages. Working on that assumption thus requires research development to occur before 2020, therefore meaning that there is some time before China can begin its period of transformation. Further, even if the plan begins on time in 2020, the period will also be burdened by social and economic change because of AI implementation.
Factors to Consider
There are more factors at play than simply putting a focus on AI. Talent acquisition, investment, research funding and cultivating legal support are all issues that China will need to resolve in order to put the plan into optimal implementation.
Talent acquisition, for example, is a key component, China has sought to draw upon world leading talent in the ‘Thousand Talents’ plan. The plan, seeking to remove hindrances and blunt shortcomings of China’s current capacity to field AI innovation, recognises a gap between China and other advanced countries, calling for better cooperation between companies and universities, leading researchers and overseas talent.
Further, the amount of data that can be produced, potentially for millions of consumers, through the use of AI systems has implications for cyber security and data confidentiality. How the Chinese Government reconciles a free and open environment for AI development with the latest round of tough cybersecurity regulations has yet to be seen, particularly with the requirement that all data collected in China must remain in China.
How China Can Achieve This
China certainly has the human capital and resources available to manufacture an environment to catalyse AI development, but one must approach this strategically.
Market drive creates new opportunities which are then filled by AI, this develops, creating new applications and then the market drives value. The circle then repeats. Should China simply command that advancements be made, one runs the risk of stifling innovation through expectation, with overinvestment, oversupply and skewed incentives destroying value in the end product. Necessity is the mother of invention and market drive is the force in this instance.
Thus, a key strategic element in developing AI is ensuring the market drive is there, in the form of encouraging broader adoption of AI within traditional industries. AI can deliver impressive competitive advantages, particularly if adopted early and early adopters are more likely to become serial adopters according to McKinsey. To take Amazon as an example, the $775m acquisition of Kiva, a robotics firm specialising in automating packaging, reduced packing time to 15 minutes, from the human worker equivalent at 60 to 75 minutes, increasing inventory capacity by 50% and reducing operating costs by 20%. The greatest strength in China’s plan is the economic potential of AI to revolutionise traditional industries.
Companies of all sizes rely on forecasting demand, seeking to optimise their supply and workforce, thus being able to reliably forecast is creating demand for AI to digest large datasets, plot trends and react accordingly. Minimising waste, augmenting strategy and policy creation as well as predicting future trends can help provide businesses with key information to maximise profit. An industry example would by Chinese healthcare, being able to predict epidemics as well as ensure medications are supplied accurately would benefit all citizens, having a public element to AI development. This can extend to utilities, optimising electric tariffing for example. AI forecasting is already well documented in retail and China possesses the opportunity to push the boundary, intriguing other industries to adopt.
A hurdle to overcome is perception, Chinese tradition is draped in history and meaning thus manufacturing a sense of urgency in adoption to bring industries forward is necessary. According to The rise of the machines: How Chinese executives think about developments in artificial intelligence, McKinsey & Company, December 2016, more than 40% of Chinese Companies do not have AI as a strategic priority. For example, Chinese agricultural firms rarely record data on planting patterns, weather impact or forecasted seasonal changes, compared to the UK, Japan and the US which have nationwide systems to capture such data and supply advance policy on these issues. Combating resistance to adoption is key in driving innovation forward.
Policies and Guidelines
To usher in adoption, providing incentives to adopt is one strategy, using legal frameworks to remove legal uncertainties, from employment issues to definitions of work procedures, can help shift this balance. Using definitive policies and guidelines on what can and cannot be done with data speeds up progress, adoption and drives that market value. One would only need a handful of competitors to take up AI practices before one feels the pressure and follows suit. To this end, China can foster industrial co-operation as standardisation in practices obliges consultation with many industry leaders about the way they use data, allowing the government to approach key stakeholders and tweak the legal environment to ease adoption and progress.
Further, China can instigate international cooperation as AI has the potential to add to China’s recent political moves, including joining the World Economic Forum, allowing greater cooperation in development. Joining a global network and hosting AI events, such as the China-Britain Business Fusion (CBBF) China-Britain AI Summit 2017 taking place in September.
Another strategic priority is to ensure that there is a ‘robust data ecosystem’ to not only control how data is used but to ensure that talent is attracted to the abundance of easily usable data. AI requires large sets of big data to learn and advance, thus being able to access data sets easily and without hindrance is a strong target. Implementing data standards, from collection and storage to security requirements can ease the process of creating a vast network of data.
Within this, standardising data sharing will increase operability between systems, allowing multiple AI to use similar data for very different purposes without having to restructure one’s own operations, from advertisers to driverless cars, understanding travel data of consumers can provide an edge. China is, as McKinsey highlight in their Artificial Intelligence: Implications for China report (April 2017), uniquely positioned in this regard, given the nation’s unrivalled ability to create data thanks to its 700m+ smartphone users.
To create a standardisation movement, the government can open more datasets to the public as well as ensuring that companies store data in certain formats or use specific procedures for storing data. Naturally, if the standardisation moves from storage procedures, then international companies will also be required to follow the regulation, allowing for even more standardised data. Whether or not this is desired is another question. A strong data infrastructure and ecosystem will also contribute to talent acquisition.
This leads into a further strategic priority, creating circumstances that are attractive to talent, whether that is domestic, incubated talent or overseas acquisition. To truly succeed in in its vision, China needs to have elite scientists and leaders pushing the boundaries of innovation and development, assisted by the up and coming AI thought leaders. There is no point in building a legacy if there is no one to inherit it. University programmes, promoting ties between industry and education, has long-term benefits for talent health, with Google’s and Microsoft’s investments in AI labs at the University of Montreal serving as an example. Abroad, China should seek top international experts in the field, opening itself up to adopting foreign academic practices and resilient methods of AI research. Standing on the shoulders of giants will go a long way in fulfilling the vision.
Hailed as the Google of China, Baidu’s latest AI developments are showing the Chinese firms can play ball with Western giants. At a conference on July 5th 2017, Baidu founder and CEO Robin Li, stated that China must remain open and continue engaging with partners around the world. Li demonstrated strides in driverless car technology by live broadcasting his commute in two driverless cars, which went without an issue. Comically, this prompted Beijing Traffic Authorities to investigate the matter as current laws do not permit autonomous vehicles on public roads.
Further, Baidu is using AI to optimise search results and build augmented reality tools. Baidu’s parent, Alibaba, accounted for 29% of China’s $42bn digital-advertising market in 2016. Baidu’s AI task force consists of 1,700 members working in 4 AI labs in China and Silicon Valley with a second lab in the Valley ready to add a further 150 scientist. Data collection and AI research are Baidu’s strengths, with 665 million monthly active users, Baidu’s AI will have no shortage of training data. This data is also unique, as compared to Google’s reach in the US and Europe.
The data originates from users who share the same culture, legal obedience, language and nationality. Thus, this results in the data being highly targeted and nuanced, allowing an AI to be created specific to the Chinese market and with such volumes of highly-specific data, the AI will equally show nuanced results and progress. Baidu’s DuerOS, conversation based AI technology set to be used in cars, home appliances and wearables, the system is already popular with NVIDIA, Haier and HTC, which, alongside Baidu’s car AI Apollo 1.0, has shown Baidu means business.
Baidu has also had its fair share of acquisitions, with Raven Tech in February 2017. This small start-up specialised in home AI, allowing Baidu to supercharge its assistant, LittleFish, contending with Google’s and Apple’s home AI to be a top performer. KITT.AI, a company specialising in natural language and speech technologies, alongside xPerception, a US-based self-driving and augmented reality technology company, will help push Baidu’s AI efforts in the coming years. Partnering with SAIC to connect cars to the internet, using cloud technology to enhance navigation functions with also promote Baidu’s effort in the autonomous car market.
A relative late-comer to the AI Party, Tencent, unveiled an AI voice platform named Xiaowei, using WeChat’s voice recognition technology. Whilst late to the gathering, Tencent again has another advantage that Google may not, a sprawling ecosystem of common users. WeChat has over 900 million users in China, who are, as aforementioned, using the same language to send voice recordings and use voice commands on WeChat. Whilst WeChat presents some difficulty to Apple’s iOS advantage due to its vast array of apps and pay function, the value for AI research is in its sheer volume of data that is collected.
Asus plans to use Xiaowei to launch educational robots, whilst WeChat will continue to seek partnerships in games, applications, news platforms and other vital functions.
China’s AI agenda and ambition for 2030 is a mix of patriotism, pride and hopefulness. To succeed, China will need to strategically deploy its human capital and change the legal environment to fully exploit AI talent. Alongside the One belt One road, China could expand this ambition to cover other countries and organically grow a network of AI partners, in the likes of India. This initiative has full state support, with a large capital backing, it is a colossal, ground-breaking initiative but it could suffer from overinvestment and oversupply. How will Chinese firms react? Is this a cause for concern for Western companies and AI leaders? That is what will be explored next.