AI Ambitions: Challenges to Innovation as the US and China Drive Global Tech
AI Ambitions: Challenges to Innovation as the US and China Drive Global Tech
AI Ambitions: Challenges to Innovation as the US and China Drive Global Techhttps://www.blackpeakgroup.com/wp-content/uploads/2019/03/Blackpeak-Insights-AI-Ambitions-Challenges-to-Innovation-as-the-US-and-China-Drive-Global-Tech.jpg1200675BlackpeakBlackpeakhttps://www.blackpeakgroup.com/wp-content/uploads/2019/03/Blackpeak-Insights-AI-Ambitions-Challenges-to-Innovation-as-the-US-and-China-Drive-Global-Tech.jpg
Both despite and because of the US and China competing in AI innovation for geopolitical and economic reasons, the two countries contribute to the industry’s development with their different tech and policy orientations. However, each approach faces unique challenges.
The Competitive Synergy of AI
The US and China dominate the global artificial intelligence (AI) industry. There are about 4,500 AI companies, around half of which operate in the US and a third of which operate in China. The top companies that employ AI talent are all either American or Chinese: Google, Facebook, Microsoft, Amazon, Baidu, Alibaba, and Tencent. The US has the most AI talent and the highest number of AI companies, but China ranks number one for quantity of AI patents, papers, and funding.
Although these countries are in their own league, the so-called “AI superpowers” have different tech advantages in the ongoing AI race. While the US has been better at designing the AI technique of deep learning, China has been better at implementation. Chinese AI technology often uses American chips from companies such as Qualcomm and US software such as TensorFlow, but China actually uses more AI than the US does. Meanwhile, China’s larger population and more data per user – thanks to mobile phone payments – give China more raw data than any other country. According to the eminent AI expert and venture capitalist Kai-Fu Lee, “In the age of AI, data is the new oil, so China is the new Saudi Arabia.” This wealth of big data translates into a competitive edge for using AI nationally.
In February 2019, US President Donald Trump signed an executive order that directs the US government to prioritize and work with industry to support the national development of AI technology. This executive order is a response to growing concerns that China’s AI capabilities have been catching up to those of the US. The rise of a serious strategic rival in an industry with clear national security and economic considerations may be forcing the US to shake up assumptions about the role of government in tech. The net result is that the US and China are increasingly aligned in the view that government should promote – or even guide – innovation, even if in reality a significant gap remains between how much each government works with industry.
Partly born of national competition, these complementary strengths between organic AI design by the American private sector and ambitious country-wide AI implementation by the Chinese government create an innovative synergy that stimulates global AI tech.
Political Risks to the AI Industry in China and the US
In 2017, China’s Next Generation Artificial Intelligence Development Plan (新一代人工智能发展规划) called for the country to become the world leader in AI innovation by 2030. As with any industry related to “core” national goals and interests, there are huge market opportunities if a company is able to survive market competition. In China, that investment is being avidly pursued by both the public and private sectors. It is estimated that up to 85% of Chinese companies are already successfully adopting or testing at least some AI. This level of implementation is much higher than those of the other countries surveyed – including the US, which came in at a distant second with 51%. Meanwhile, China also has more than a dozen AI companies worth at least USD 1 billion.
Yet this stamina for AI advancement has led to concerns that China’s AI industry might be facing a bubble. In 2017, Chinese companies received 70% of global AI investment, but the tide of easy funding may be starting to recede as the industry grapples with commercializing the technology. The appeal of China’s AI tech may also be limited outside the country, as a mere 4% of Chinese AI patents are later filed in other jurisdictions, compared to 32% of US AI patents. In fact, researchers and policymakers in China have warned local governments and investors not to get too carried away by AI mania. A Tsinghua University report on the Chinese AI industry pointed out that some local governments have been “blindly following the central government’s AI policy” and failing to take local context into consideration when developing AI. Time will tell how much of China’s enthusiasm for AI’s potential can translate into sustainable development and profits.
Changing Data Privacy and Cybersecurity Regulations
Key to the development of AI is data: the more, the better. At the national level, China’s data privacy regime has been largely fragmented among assorted laws, measures and regulations. This lack of an overarching regulatory framework has allowed Chinese apps to access vast quantities of user data, which has generally not caused public concern. Baidu CEO Li Yanhong’s comment that Chinese users care more about convenience than privacy is a widely held view in the tech industry. However, consumer backlash has pushed the issue to the fore. Most notably, in January 2018, Alibaba affiliate Ant Financial had to apologize publicly for automatically signing up users for a social credit program without properly obtaining their consent. This free reign that companies previously had to take advantage of the loose policy environment may be coming to an end.
From mid-2017 onwards there has been a marked shift both in policy and public perceptions on the importance of data privacy. On 1 July 2017, China rolled out its Cybersecurity Law (网络安全法), which was the first to introduce a comprehensive set of data protection provisions. Less than a year later on 1 May 2018, China issued the Personal Information Security Specification (个人信息安全规范) containing detailed guidelines for data handling and data protection. Similarly, in recent years the Chinese government has circulated hundreds of new national cybersecurity standards relating to various software and hardware products. Some of these standards can pose challenges to foreign firms operating in China because “invasive product reviews”of foreign companies’ intellectual property could be mandatory and products may have to be redesigned from international standards to meet the new local standards. As the policy environment is in transition, companies will have to adapt to the stricter data privacy policies and foreign firms will need to cope with the new cybersecurity standards.
Lack of US Policy for AI May Hinder Competition by AI Start-Ups
Machine learning – a common AI technique – and big data further intensify network effects, which make it increasingly difficult for new firms to enter the market. At the same time, it is not uncommon for major tech companies to acquire AI start-ups that are successful. One US senator recently suggested policy measures such as legislation to increase data transparency and decrease data portability (the consumer cost of switching to new digital providers) to blunt the AI-enhanced network effect advantages of the large digital companies. For now, however, the lack of AI policy may harm competition in the US tech industry, which already has only a 65% concentration of AI usage. In contrast, Chinese companies’ AI usage is consistently above 80% across both the tech and other industries.
Disconnect Between Public and Private Sectors Causes Uncertainty
How the giants of Silicon Valley respond to President Trump’s executive order has wide ramifications for the future of AI development. So far, large companies have had strikingly different approaches to working with the US government on AI projects. For example, Microsoft, Amazon, and others have been fiercely competing in an ongoing bid for the $10 billion contract for the Pentagon’s Joint Enterprise Defense Initiative (JEDI) program, a Department of Defense project to use cloud computing and facilitate the usage of AI on the battlefield. Google, on the other hand, refused to bid on the project. In June 2018, the search engine company had even canceled its participation in the Maven program, a joint AI project with the Pentagon, over ethical concerns about the potential lethal usage by drones of the video-processing technology.
Analysts worry that without more public-private AI cooperation, the US may not be able to maintain its technological edge. Although a policy for more collaboration between government and industry was a leading element of the executive order, details on how to achieve this and other AI policies were sparse. At the same time, leaders in Silicon Valley are more likely than their Washington counterparts to view China in business, rather than geopolitical, terms. It remains to be seen whether President Trump’s executive order will help bridge some of the cultural and philosophical divide between the American political and tech capitals.
Thinking About Global AI Strategies
Official government strategy is a relative latecomer to AI innovation. AI tech has been around in various forms for decades, while national strategies have started to spring up over the past 2–3 years. In this new era of government-led AI initiatives, the US and China represent two different models for national AI development. Will other countries adopt the policy-light, bottom-up approach to innovation seen in the US? Or will governments around the world increasingly formulate and implement policy to catalyze AI advancement?
So far, several major AI players – both in the West and elsewhere – have started to have a similar approach as China’s. Although countries as geographically diverse as Canada, France, and Singapore might emphasize different paths to national AI development, they agree that government initiatives can be a driver of the process. As AI tech continues to evolve, investors should remain well-informed about how both the presence and absence of AI policy can present challenges to the industry of the future.