How Countries Are Preparing for the Global AI Race

Like the introduction of other major technologies, from electricity to the Internet, the mass adoption of AI has the opportunity to substantially grow and disrupt the global economy. The extent to which governments realize this and take action to support the growth of AI will play a large rule in determining their future economic competitiveness and quality of life for their citizens. On July 25, 2018, the Center for Data Innovation held a discussion on the various strategies countries are using to promote the development and adoption of AI. Center for Data Innovation senior policy analyst Joshua New moderated the conversation with experts representing China, Germany, India, and the United Kingdom.

Xiaomeng Lu, international public policy manager at Access Partnership, described China’s ambitious goals for AI development over the next 12 years which is to become the global leader. China wants to be equal to countries leading in AI by 2020, focus on developing breakthroughs in areas of AI with high economic value through 2025, and by 2030, become the worlds “premier artificial intelligence innovation center,” with a domestic AI industry worth approximately $150 billion. Despite these lofty goals, Lu noted that China faces significant obstacles in its path to AI dominance, particularly the challenges posed by China’s restrictions on the free flow of data across its borders and between the public and private sectors, which limits the quantity and quality of data available to spur innovation in AI.

Germany, which is still developing its national AI strategy, is taking a different approach from China, opting to capitalize on its existing strengths and target specific weaknesses to improve its competitiveness, rather than aim to be the world leader in AI. Robin Mishra, head of section science and technology at the Germany embassy, explained that while Germany has traditionally been very successful in fundamental scientific research and development (R&D), it struggles to translate these efforts into applied R&D and share these innovations with the private sector. Thus, Mishra highlighted how improving technology transfer to the private sector will be a key focus of Germany’s forthcoming AI strategy. Mishra also described Germany’s plans to develop the human capital necessary for AI, including by increasing opportunities for AI research in academia, integrating AI skills into college degrees beyond just those in science, technology, engineering, and mathematics (STEM) fields. While Lu lamented China’s reluctance to provide the data necessary for the private sector to develop AI, Mishra described how Germany intends to develop a more comprehensive open data strategy that could provide valuable data to AI researchers.

Like Germany, India is still developing its AI strategy, however Arunish Chawla, economic minister at the Indian Embassy, described how India’s main goals with its strategy are to support the private sector in developing AI while ensuring AI both benefits society and that the Indian economy is resilient enough to withstand the disruption it will cause. In particular, Chawla emphasized the need for India to support workforce development so that Indian workers can respond to the job displacement of widespread automation. Chawla envisioned India adopting worker apprenticeship models common in Germany to help workers learn new skills, and both he and Lu stressed the value in developing “soft skills,” or skills not directly related to AI like programming or data science, as not every person in an AI driven-economy will need to be a coder. Chawla also described how India’s plans for AI are unique in that they heavily emphasize the importance of harnessing AI to deliver social benefits alongside economic benefits.

Andrew Price, head of the U.K Science and Innovation Network, Americas, for the British Embassy described how the United Kingdom’s AI strategy, called the AI Sector Deal, has a large focus on data—particularly data interoperability between industries and more open data standards. Price pointed to the development of data trusts—frameworks for sharing proprietary and sensitive data between government, industry, and academia with clearly defined rules and responsibilities—as a promising example of how the public and private sectors could collaborate to advance AI research without sacrificing privacy or security. Price also described the new Centre for Data Ethics and Innovation, established as part of the AI Sector Deal, which will develop methods for measuring and improving trust in data-driven technologies like AI, as well as recommend ways for addressing the novel ethical challenges AI could pose.

All panelists agreed on the importance of carefully monitoring the consumer protection issues related to AI, though they described that their respective countries were all approaching the issue somewhat differently. Lu, not an official representative of China, explained that China will likely not place much emphasis on ethics and privacy concerns initially. Price agreed that AI will pose new challenges, but believed that regulating too soon could be harmful, and that governments should take sector-specific approaches to regulation. Chawla expressed support for international collaboration around developing ethical standards and rules for the use of AI, however Mishra and Price argued this too would be more effective as a sector-specific approach, such as international rules about the use of autonomous weapons, rather than rules governing the use of AI generally.

As the United States does not yet have a national AI strategy, the panelists offered advice based on their experiences with their respective countries’ efforts. Lu and Mishra recommended the United States focus on continuing to enable the private sector’s development of AI to capitalize on what has long been its main strength in innovation, and avoid the central planning-style approach of China. Chawla stressed that as a leader in AI, the United States has a responsibility to lead international forums in shaping the technology so that it benefits global value chains and the good of humanity as a whole. Finally, Price recommended that the U.S. AI strategy target more than just the public footprint of AI—i.e. public sector R&D—and that the whole of government should be involved in the development and execution of a national AI strategy, rather than just scientific agencies.