Yesterday, ITIF’s Center for Data Innovation held an event titled, “How Countries are Preparing for the Global AI Race.” The panelists touched on the approaches that the United States, United Kingdom, Germany, China, and India are taking with respect to the development of AI. As the panel progressed, it became clearer that there is no singular “key” to “winning” the global AI race.
It is important to recognize that the challenges in AI advancement vary from country to country. For example, China has a competitive advantage in terms of data, but it still has a considerable skills gap that is hindering that data from being utilized in an AI capacity. According to panelist Robin Mishra of the German Embassy, Germany has a highly-skilled labor force and has invested a considerable amount in research and development, yet it lacks robust industry outside of manufacturing that can take advantage of further developing AI in the way that the United States or China can with their tech sectors. As such, it’s difficult to pinpoint one specific area or condition that is key to a country dominating in AI as is true with many other technologies.
France and China released their own roadmaps for AI development, setting goals and committing to AI initiatives. Some countries have even released their own sets of AI principles. However, the notion that each country needs to have a comprehensive AI strategy seemed to be disputed among the panelists.
Panelist Arunish Chawla of the Indian Embassy stated that the reason why India became a telecommunications superpower was precisely due to the lack of a top-down policy dictating the path forward for Indian tech companies and setting rules in place. He stated that India intends to take the same approach with respect to AI. Panelist Andrew Price of the British Embassy was in agreement with the Indian panelist about not needing a top-down policy strategy for a country to be successful. Even panelist Xiaomeng Lu of Access Partnership said that the Chinese government has set certain benchmarks that will be difficult to meet if additional steps are not taken. Instead policymakers should be guided by the private sector to find areas of need and improvement and act to fix those areas to facilitate innovation.
Given these challenges, it begs the question of what the role of policy and government should be in the development of AI. The panelists named quite a few areas where they were in agreement that government can actually be helpful.
First, governments can ensure success in the AI space is by investing in AI research and development. This can help development in emerging and promising technologies. Much like with the development of GPS, public investment in R&D can also help achieve insights that can then potentially be used for commercial gain. It can also produce valuable public data sets that can then be used by anyone seeking to gain entry into the AI space. Germany, France, and China have committed to providing such incentives and to invest more in AI R&D. The Trump Administration expressed support for doing so recently in “Artificial Intelligence for the American People” – the prioritized funding for AI is highly welcome.
Policymakers can also invest in and implement training and retraining programs for workers in order to fill the skills gap is also critical. According to the panelists, it is important that this is done in partnership with private companies so that any technical programs actually teach the skills that employers need filled. It doesn’t matter how much money countries invest in AI or how much data a country possesses if there aren’t enough workers to make use of the data. As with other emerging technologies, workers need to be incentivized, trained, and retrained in the necessary skills to fill the plentiful jobs that will be created as a result of AI.
Finally, panelists also stated that governments should work with the private sector and other stakeholder groups to establish a set of best practices and set the rules of the road for approaching AI. Carefully crafted AI principles for social good will serve to alleviate public concerns about the negative impacts that AI may have on humanity, providing an avenue for accountability. If developed in tandem with stakeholder groups, such principles can be developed so as not to hinder innovation.
These suggestions and approaches are not new. SIIA identified many of these issue areas as well. SIIA has even released its own ethical principles for AI and a brief on the future of work. Perhaps the biggest takeaway from this panel is that policy is not the singular answer to a country becoming a leader in AI, nor is solely investing in R&D. Policy can be most successful in this space if governments use it to assess their own areas of need and implement programs to address these needs. Moreover, governments can work internationally to promote AI. The G7 “Charlevoix Common Vision for the Future of Artificial Intelligence” provides an excellent roadmap, which future G7 and perhaps G20 summits can build upon. The emphasis in that document on cross-border data flows and interoperability is an important element in realizing the full potential of AI.