With the signing of an executive order, the American AI Initiative, President Trump signaled on Monday that the U.S. would not sit back and let China take the lead on artificial intelligence. That’s a great first step, but the U.S. faces a tough road ahead in keeping its technological lead.
And maintaining that lead should be a top priority. Not only does AI have military applications critical to ensuring the country has robust national security (Skynet anyone?), but it is likely to be the key economic driver of the not so distant future. Indeed everything from automated cars, to fraud detection in financial transactions, to translation, and even farming as well as other applications in most every other industry already heavily rely on machine learning.
If the U.S. fails to develop its AI capacity, it risks being entirely dependent on China for key parts of military infrastructure as well as lagging in its ability to develop and integrate new technology into its economy.
The order calls for more federal investment in AI research, to set standards, and jump-start training necessary for AI jobs among other provisions. It is, however, more aspirational than comprehensive. It allocates no new funds for AI and lays out no framework for its key components. Still, it sends a clear message: The U.S. has AI on its radar.
Although it’s an important step, the U.S. effort pales in comparison to China’s. In July 2017 China’s State Council released the “ New Generation Artificial Intelligence Development Plan” with the ambitious goal of becoming an AI power by 2030. That plan fits well into the country’s larger goals, including the “ Made in China 2025” initiative that aims to root the development of future technologies in China, rather than the West.
Both of those initiatives, and the nature of China’s state-led economy, come with substantial investment at both the national and local level as well as clear, if daunting, goals and a substantial development plan. That’s tough competition for Washington.
Trump’s order will, however, take on some of these challenges.
For one thing, the success of AI development depends on large data sets. And in the U.S., unlike China where there are much stronger ties between private companies and the government, the government and private industries have been reluctant to collaborate over privacy and ethics concerns.
This was clearly on display in 2018 when Google employees protested the company’s collaboration with the Pentagon, forcing the company to drop its partnership over concerns over lethal uses of AI developments.
Other efforts, however, have had more potential. A project to share anonymized veteran’s heath data with Google’s parent company Alphabet to use artificial intelligence to predict the development of kidney damage, for example, could save lives and serve as a model for other data sharing without privacy violations.
Indeed, Trump’s executive order outlines the potential for more collaborative sharing of certain data between the government and researchers and private companies.
Additionally, political opposition to AI in the U.S. often centers on the impact of new technology on jobs and communities already hit by the loss of American manufacturing. That’s undeniably true, and a transition to jobs compatible with more automation will not be easy.
Trump’s plan deals with this by explicitly mentioning jobs training and incorporating that from the very beginning into his plan for AI. Without funding, of course, those mentions are merely words, but it’s great to see that incorporated in from the get-go as it will, hopefully, lessen political opposition.
Trump has a long way to go on AI, but this executive order is an important step in ensuring that the U.S. remains the economic and innovative powerhouse. Hopefully, this is just the start of engagement with technology that will likely redefine the country’s economic landscape.
window.fbAsyncInit = function() FB.init(
appId : '190451957673826',
xfbml : true, version : 'v2.9' ); ;
(function(d, s, id) var js, fjs = d.getElementsByTagName(s); if (d.getElementById(id)) return; js = d.createElement(s); js.id = id; js.src = "http://connect.facebook.net/en_US/sdk.js"; fjs.parentNode.insertBefore(js, fjs); (document, 'script', 'facebook-jssdk'));