Encouraging Cross-Border Data Flows Can Enable the U.S. to Maintain its Edge in AI

Global Connections (World map texture credits to NASA) Photo Credit: Getty Images.


By: Diane Pinto, Columnist

When it comes to artificial intelligence as with other commercial and military tools, it is strategically best for U.S. interests for the U.S. to remain at the forefront. So, when reports come out that China could overtake the U.S. due to the amount of data it has, largely due to the sheer size of its population, it makes sense that U.S. policymakers are concerned.[i] Data is the foundation for all forms of software, and the development and application of artificial intelligence (AI), in comparison to traditional forms of software, requires far more vast quantities of diverse data to yield accurate, meaningful results. This is especially true when venturing into subsets of AI applications like machine and deep learning where algorithms learn and train themselves based on data inputs. But, much like with more traditional software, developing AI tools and having the ability to utilize them is largely reliant on international trade mechanisms since AI technology often draws from data controlled by several countries.[ii] Therefore, while the U.S. might not be able to produce data in the same volume as China can, it can still maintain superiority by encouraging a cooperative international environment that promotes cross-border data flows. By doing this, the U.S. can remain competitive not just commercially, but for national security purposes as well.

AI tools are used for a wide variety of national security-related purposes and will only proliferate to make tasks easier for humans. This is to the strategic benefit of the U.S. as it constantly seeks technological superiority over its adversaries. For example, language and content data is needed for AI-enabled anti-propaganda and anti-disinformation tools to interpret sentiment and filter out fake news.[iii] Financial transaction data is needed by AI tools used for pattern recognition and verification to counter terrorist financing and money laundering.[iv] Integrating AI into existing weapons platforms and new ones can improve conflict decision-making, precision-targeting of weapons, quality of training and wargaming, logistics, and cost-savings measured in both blood and treasure.[v] Because the sources of the data needed for these types of AI-tools often span several countries, data needs to flow across borders as unhindered as possible for access and utilization. This is best achieved through international trade mechanisms to balance legitimate concerns like access to sensitive information with legitimate national security needs.

There are several challenges facing cross-border data flows that, if not dealt with correctly, will have negative consequences for AI development as well. The competition for data control may lead some countries to instill data protectionist policies. Data localization measures require companies to store data in centers within their borders.[vi] These requirements place undue burdens on U.S. companies while also raising data security concerns. Localization requirements, enacted by several countries, account for significant losses in GDP. In China alone, data localization has been measured to account for a loss in its GDP of roughly 1.1 percent.[vii] If more countries continue to enact localization requirements, U.S. firms developing AI will be challenged and have to make decisions about where they should and should not operate, antithetical to the optimal conditions needed for producing well-functioning AI technologies. The U.S. should dissuade countries from enacting such requirements that are not just bad for it, but cause self-inflicting damage for the countries implementing them.

Conversely, there are positive ways to address legitimate concerns about cross-border data flows that can still be conducive for AI development. Data privacy is one of these legitimate concerns, and it continues to be a hotly debated topic among policymakers as they try to reconcile the need for free-flowing data with consumer privacy and speech rights. Fortunately, there are several positive examples of international bodies facilitating data flows that also respect privacy. First, the Asia-Pacific Economic Cooperation Cross-Border Privacy Rules system facilitates data flows while promoting privacy protection by requiring that countries who accede to the system comply with the APEC Privacy Framework.[viii] Next, the EU-U.S. Privacy Shield provides companies with low-cost means of moving data between the U.S. and the EU provided that companies who self-certify commit to complying with the Privacy Shield’s consumer notice, data protection, and data retention requirements.[ix] Further, the U.S. Mexico Canada Agreement contains a digital trade chapter that includes provisions banning data localization, encouraging the free flow of data, and promoting data privacy.[x] Systems like these can serve to facilitate an environment for the U.S. to thrive as it continues to develop AI while also accounting for legitimate concerns like privacy.

As battlefields evolve, having access to several kinds of data – like terrain, intelligence, or social media data – that flows easily across borders will ensure that AI-enabled platforms can perform up to task in different types of conflict scenarios that humans might not account for on their own. To make up for any dearth of data the it might have access to in its quest for AI superiority, it is critical that the U.S. encourages an international system that facilitates cross-border data flows. It should discourage countries from implementing legal frameworks that hinder data flows as they have demonstrably negative consequences. Otherwise, the scope of data inputs will be limited leading to inaccurate results or have negative consequences like inaccurate targeting or failing to flag several social media pages as propaganda machines or terrorist recruitment vehicles.

The U.S. should continue to strive for an international regime that promotes cross-border data flows while accounting for legitimate concerns like privacy and speech rights like it has in both the EU-U.S. Privacy Shield and the U.S. Mexico Canada Agreement. It should seek to include such provisions in future trade agreements especially as it looks to negotiate further free trade agreements with the U.K., EU, Japan, and more countries.[xi] By encouraging an international system that fosters cross-border data flows, the U.S. can still maintain access to other sources of data across borders to ensure its AI-enabled national security tools remain at the forefront even though the U.S. on its own might not be able to produce as much data as does China.

Bibliography

[i] “The algorithm kingdom: China may match or beat America in AI.” The Economist. https://www.economist.com/business/2017/07/15/china-may-match-or-beat-america-in-ai, July 15, 2017.

[ii] Susan Aaronson. “Artificial Intelligence is Trade Policy’s New Frontier.” Centre for International Governance Innovation. https://www.cigionline.org/articles/artificial-intelligence-trade-policys-new-frontier, January 11, 2018.

[iii] Michael Horowitz, Paul Scharre, Gregory C. Allen, Kara Frederick, Anthony Cho, and Edoardo Saravalle. “Artificial Intelligence and International Security.” Center for a New American Security. https://www.cnas.org/publications/reports/artificial-intelligence-and-international-security, July 10, 2018.

[iv] Ibid.

[v] Robert Warren Button. “Artificial Intelligence and the Military.” Rand Corporation. https://www.rand.org/blog/2017/09/artificial-intelligence-and-the-military.html, September 7, 2017.

[vi] Erik van der Marel, Hosuk Lee-Makiyama, and Matthias Bauer. “The Costs of Data Localisation: A Friendly Fire on Economic Recovery.” European Centre for International Political Economy. http://ecipe.org/publications/dataloc/, May 2014.

[vii] Ibid.

[viii] “About CBPRs.” Asia-Pacific Economic Cooperation. http://cbprs.org/about-cbprs/, 2019.

[ix] “The EU-U.S. Privacy Shield Framework Principles.” U.S. Department of Commerce. https://www.privacyshield.gov/EU-US-Framework.

[x] Chapter 19: Digital Trade. United States Mexico Canada Agreement. Office of the U.S. Trade Representative. https://ustr.gov/sites/default/files/files/agreements/FTA/USMCA/Text/19_Digital_Trade.pdf.

[xi] “Trump Administration Announces Intent to Negotiate Trade Agreements with Japan, the European Union, and the United Kingdom.” Office of the U.S. Trade Representative. https://ustr.gov/about-us/policy-offices/press-office/press-releases/2018/october/trump-administration-announces, October 16, 2018.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.