The US is finally moving towards an AI strategy

The US is finally moving towards an AI strategy
© Getty Images

Many have called on Washington to develop a national strategy for artificial intelligence (AI). On Feb. 11, the Trump administration showed that it was listening. President TrumpDonald John TrumpCummings says Ivanka Trump not preserving all official communications Property is a fundamental right that is now being threatened 25 states could see severe flooding in coming weeks, scientists say MORE issued an executive order to establish the “American AI Initiative,” which seeks to improve AI R&D, workforce development, and international engagement. The American AI Initiative is a welcome step towards boosting U.S. competitiveness in AI, but it does not go far enough. Without more substantial efforts from the Trump administration and Congress, the United States will struggle to compete with other countries that have developed aggressive national AI strategies, and the economy will be worse off for it.

The American AI initiative directs federal agencies to take five important steps to increase U.S. leadership in AI: prioritize AI in their R&D missions; make resources critical to AI development, including data and computing resources, available to researchers and industry; establish guidance and standards for AI in various sectors; prioritize worker training programs to help American workers develop skills necessary to work with AI; and the administration commits to ensuring that international markets are open and competitive, as well as conducive to AI R&D.

ADVERTISEMENT

However, what the American AI Initiative does not do is noteworthy. Despite its heavy emphasis on R&D, computing resources, and skills development, Trump’s executive order simply reprograms existing funds for these efforts, rather than significantly increasing investment. Instead of taking direct action, in most cases the executive order simply directs various agencies to develop recommendations for action within 180 days. Neither does the executive order encourage public sector adoption of AI or call for light-touch regulation of AI, both of which would be a major boon to U.S. leadership in AI.

In contrast,  Canada, China, France, the United Kingdom, and many other nations already have comprehensive national AI strategies that take much more decisive action to support AI and support these initiatives with generous funding.

The American AI Initiative can serve as the foundation of a national strategy, but Congress and the administration must build on this foundation to be effective. There are several key areas that would benefit from swift action. First, policymakers should substantially increase AI R&D investments, which will be a critical factor in determining the potential for U.S. firms and researchers to develop innovative new AI applications.

Second, policymakers should prioritize cultivating high-skilled AI talent in higher education as well as implement comprehensive reforms to the United States’ workforce training and adjustment policies. While the private sector invests in worker training, it does not do so at optimal levels because it cannot capture all of the benefits of doing so. For example, it is more cost effective for a firm to hire high-skilled AI talent away from competing firms than to invest in long-term worker training for its own employees, causing all firms to underinvest in skills development. This also leaves firms that cannot offer competitive salaries for AI talent unable to take advantage of AI, despite the benefits it could provide.

Third, policymakers should enable the public sector to more quickly adopt AI. Despite the benefits AI could offer the public sector, federal agencies face significant challenges when it comes to using the technology, such as a lack of awareness and strategic leadership, that require policymaker intervention to address. For example, to jumpstart this process, the president could direct every federal agency to identify and launch at least two high-impact AI projects within the next 12 months. If policymakers, particularly in the executive branch, were to provide this much-needed guidance to spur public sector adoption of AI, the benefits to federal agency missions and to taxpayers would be immediate and substantial.

ADVERTISEMENT

Fourth, policymakers should do more to provide the technical resources necessary for AI development, especially data. For example, federal agencies should support the development of shared pools of high quality, application-specific training and validation data in key areas of public interest, such as agriculture, education, health care, public safety and law enforcement, and transportation. One way to do this is for agencies to create “data trusts” that have clearly defined rights and responsibilities for how different parties can use pooled data from individuals and the private sector. Data trusts could greatly increase the amount of data available for data innovation and AI development in sectors such as health care, which often deals with sensitive information, and for the economy as a whole, as businesses would have greater confidence to share their proprietary data.

The American AI Initiative is proof that the administration recognizes the importance of the government’s role to play in advancing U.S. leadership in AI, and thus should be met with cautious optimism. Policymakers on both sides of the aisle should start building on this by crafting a more detailed national AI strategy designed to maximize the social and economic value of AI. Many other countries have already recognized supporting AI as a major policy priority and have enacted state-level AI strategies of their own to improve their technological competitiveness. It is time for policymakers in the United States to do the same.

Daniel Castro (@castrotech) is director of the Center for Data Innovation (@datainnovation), and Joshua New (@Josh_A_New) is a senior policy analyst at the Center for Data Innovation.