The American AI Initiative: What Could Implementation Look Like?
On February 11, 2019, President Trump signed an executive order dubbed the American AI Initiative which aims to outline U.S. priorities in artificial intelligence (AI) going forward. Prior to this initiative, the U.S. lacked a comprehensive national AI plan. This order follows a string of government initiatives by other countries within the last couple of years, including Singapore, Canada, the European Union, and China. The initiative listed the key areas of focus as follows: 1) investing in AI research, 2) unleashing AI resources, 3) setting AI governance standards, 4) building the AI workforce, and 5) international engagement.
Current AI initiatives can be divided into two broad groups. The first group includes specific planning and funding for AI while the second group consists of guiding documents that outline the overall direction in which the country is headed. As multiple media outlets such as Fortune, Forbes, and Bloomberg have pointed out, the American AI Initiative is limited in making specific planning commitments, placing it in the second group.
The American AI Initiative comes as a welcome starting point in the strategization of AI but inevitably raises a few questions, including: how can the U.S. transform this order from guidance to implementation, and what has the U.S. already done in respect to the order? Below are comparisons of what other countries have been doing in furtherance of the areas that President Trump’s initiative has outlined:
1) Investing in AI Research
The American AI Initiative aims to “focus[…] on maintaining our Nation’s strong, long-term emphasis on high-reward, fundamental R&D in AI by directing Federal agencies to prioritize AI investments in their R&D missions.”
Many governments have formulated substantial investment packages in their AI strategies, with numbers ranging from 21.6 million (Australia) to almost 2 billion by 2022 (South Korea). The U.S. initiative, in contrast, does not establish a definite, quantifiable commitment to spending, and total federal funding for AI can only be estimated. Individual agencies have released some numbers; the Pentagon has increased R&D spending on AI from $1.4 to $1.9 billion between 2017 and 2019, and DARPA, the Pentagon’s research arm, has pledged $2 billion in AI funding over the next five years. While Congress is the branch responsible for appropriating funding, the White House can steer departments and agencies to allocate funding without approval directly from Congress. Regardless of how the U.S. ultimately decides to implement funding, government investment in AI is critical, according to Nvidia executive Ian Buck:
“There’s simply no replacement for the federal government significantly increasing support for fundamental research to bolster university research…funding drives research. Research, in turn, drives innovation, from startups to multinationals.”
2) Unleashing AI resources
The American AI Initiative plans to “make Federal data, models, and computing resources more available to America’s AI R&D experts, researchers, and industries...”
France’s open data strategy extends the opening of datasets to the private sector too, as long as it serves the public interest. The Big Data Development Action Plan in China has been implemented since 2015 but, despite being at the forefront of the AI race, Chinese researchers still deal with a shortage of open data from the government and internationally due to firewalls and data localization requirements. Ms. Xiaomeng Lu, public policy manager for technology consultant Access Partnership, cautions against restrictions on cross-border data flows, arguing that “only the free flow of data can generate value”.
The signing of the OPEN Government Data Act in January 2019 provides meaningful advancement towards the U.S. priority to unleash AI resources. If implemented successfully, it will “transform the way the government collects, publishes, and uses non-sensitive public information,” according to Sarah Joy Hays, Acting Executive Director of the Data Coalition. One of the most important aspects of what the Act details is that it sets the presumption that all government data should be open data by default, and should be machine-readable and freely-reusable.
3) Governance Standards
The American AI Initiative says that “Federal agencies will foster public trust in AI systems by establishing guidance for AI development and use across different types of technology and industrial sectors.”
The European Union’s approach to this process has entailed establishing a new set of AI ethics guidelines to undertake issues such as fairness, safety, and transparency. To further this goal, a new High-Level Group on Artificial Intelligence will be in charge of preparing and drafting ethics guidelines for member states to consider. In Singapore, the government announced three new initiatives on AI governance and ethics.
By contrast, governance standards in the U.S. have originated with the private sector, where companies such as Google and Microsoft have established their own set of ethical frameworks. In addition, a broad number of businesses and NGOs have convened the Partnership on AI to formulate best practices. Businesses have sought input from the Oval Office, however, such as when Google released a policy paper that recommended government guidance for AI-powered products and services. These efforts signal a need for more collaboration between the government and the private sector in this area.
4) Building an AI Workforce
The American AI Initiative aims to “prepare our workforce with the skills needed to adapt and thrive in this new age of AI…[and] calls for agencies to prioritize fellowship and training programs…in computer science and other growing Science, Technology, Engineering, and Math (STEM) fields.”
In Korea, in order to secure AI talent, the government will establish six graduate schools in AI by 2022 to train 5,000 AI specialists. In Singapore, the government has partnered with DataCamp to bring the organization’s data-science, machine learning, and AI curriculum to all Singapore public schools and universities at no cost to the schools. The AI Talent Program in Taiwan aims to actively recruit globally and make it easier for individuals to start work in Taiwan. The Canadian government is investing more than $200 million in AI research programs at three Montreal universities; in 2017, it implemented a program called the Global Skills Strategy that issues temporary work permits with faster processing times for workers in certain categories, such as software engineers.
While the Trump Administration’s AI plan prioritizes the importance of building an AI workforce, efforts to do so may be hindered due to what some perceive to be increasing anti-immigrant sentiment and difficult visa policies. As a result, the pool of AI talent in the U.S. may shrink, with anecdotes of scientists and engineers in the U.S. already leaving for Canada. However, the U.S. recently made changes to its H-1B visa system, giving priority to foreign workers with advanced degrees from an American university over those hired abroad (although the time of implementation for the new system has not been determined yet). Additionally, the United States Citizenship and Immigration Services (USCIS) announced in February 2019 that it will resume premium processing for all H-1B applications filed before a certain date.
5) International Engagement
The American AI Initiative “is committed to promoting an international environment that supports AI R&D and opens markets… while also ensuring that the technology is developed in a manner consistent with our Nation’s values and interests.”
Many countries have begun a collaborative process in navigating the world of AI together. The UAE and India, for example, signed a memorandum of understanding to explore options for collaboration and foster growth in their AI economies. The UK and France are planning a digital conference together. France and Canada have announced the joint creation of a study group on inclusive and ethical AI. The U.S. does not yet have any bilateral agreements with other countries to advance and promote AI, but is working through existing NGOs to foster ethical and socially beneficial development of AI technologies.