TY - JOUR
AU - Wurst, Noelle
AB - In a scene perhaps befitting an Isaac Asimov-meets-Albert Camus short story, Amazon’s new household robot, Astro, will apparently throw itself down the stairs.1 If the artificial intelligence (AI) offered by a trillion-dollar company cannot manage the various boundaries within a home, it may seem too early to worry about AI crossing international borders. But, in fact, the world of international AI is already here. From search engines, to facial recognition, to the myriad services using cloud computing, to the tangible reality of the internet of things (IoT), AI is increasingly everywhere. Typically, these deployments of AI cross borders, implicating international economic law. Governments have taken note, with some announcing ambitious plans to regulate AI or to invest in AI, or both.2 Thus, a volume from international economic law experts analyzing AI from an array of perspectives is a welcome development. Consider just a few of the ways that AI intersects with international economic law. A country’s measures regulating AI may raise questions of compatibility with international trade law commitments if they, for example, seem to discriminate against foreign AI suppliers. Databases assembled to train AI could conceivably be treated as foreign investments subject to international investment law if they are required to be shared with competitors, as countries are now increasingly planning. Different countries may take different positions on standards for intellectual property protections for works created by AI. An AI could, for example, literally create millions of melodies and seek to copyright them. Would international intellectual property law require that those fabricated melodies receive global protection once they are copyrighted in one jurisdiction? Clever transfer pricing might enable multinational corporations to situate royalties for AI in low-tax jurisdictions; international tax law could help distribute tax payments more fairly. What international finance rules should there be for investing in AI or using AI for financial decision-making? What about export controls and sanctions for AI that might be used for weapons or surveillance?3 How should we reduce carbon emissions from AI’s massive computational requirements, implicating international environmental law? Whoever owns AI and whoever regulates it will affect economic development, inequality, and consumer protection. Shin-yi Peng (National Tsing Hua University), Ching-Fu Lin (National Tsing Hua University), and Thomas Streinz (New York University School of Law) have put together a volume of excellent papers to help us think about how to regulate the new international digital economic order that AI is creating. Most legal scholarship on AI focuses on its possible risks for entrenching bias or invading privacy or on possible uncertainties as to intellectual property rights in AI-created works. This new book expands our horizon by convening international economic law experts to study the challenge of AI. This volume widens the lens to consider issues central to international economic law—including trade, development, regulatory space, and consumer protection. The book also sheds light on some more obscure corners of international economic law—for example, the United Nations Economic Council for Europe’s Global Forum on Road Traffic Safety and standard-setting bodies such as the SAE International, ISO, and IEEE (Lin, p. 246). Ironically, the volume demonstrates the value of physical conferences, as it grew out of a 2019 Society of International Economic Law conference of the Asian International Economic Law Network. That conference, titled ‘International Trade Regime for the Data-Driven Economy: How Will Artificial Intelligence Transform International Economic Law?’, was hosted by the Institute of Law for Science and Technology at National Tsing Hua University in Taiwan. We might note that international investment law, a large component of international economic law scholarship, has perhaps thankfully, not yet taken center stage in these discussions. Perhaps, this is because we do not seem to have any invocation by investors of investor-state or by states of state-to-state dispute settlement mechanisms in investment treaties. But AI implicates foreign investment. The US government, at least briefly, sought to ban TikTok because of its foreign ownership. As of this writing, a US government divestiture order is still on the books, although it does not appear to be enforced. As Thomas Streinz writes, ‘international investment law’s bearing on data control has been largely overlooked, but this might just be the calm before the storm’ (Streinz, p. 189). He writes that it is ‘plausible that “data” will soon be recognized as a protected asset under international investment law by at least some tribunals’ (Id.). He notes that it is possible that companies may receive such property rights protection for data even when it is unavailable in domestic law (Id., n. 88.). This would, in our view, be inadvisable, unless the investment treaty made it clear that data were covered by treaty obligations. As Streinz observes, mandatory data sharing rules being considered by various jurisdictions, such as the European Union’s (EU) Data Governance Act and India’s proposed Personal Data Protection Act, will perhaps test international investment rules. In their chapter, Kelly K. Shang (University of Bern) and Rachel R. Du (University of Bern) explore trade law issues implicated in sanctions imposed either to protect national security or promote human rights. AI regulation must be understood in the context of global competition. Jane Winn’s (University of Washington) and Yi-Shyuan Chiang’s (National Tsing Hua University) paper reminds us of a growing great power rivalry in the AI space between the US and China, with both states seeking to ensure their leadership in the AI race. Alan Hervé (Sciences Po Rennes) reminds us that the European Union (EU) wishes not to be left behind. The European Commission’s proposed AI Act, for example, clearly embraces AI, preferring risk-based regulation to more prescriptive approaches. The proposal is explicit in its desire to ensure that the EU leads in AI development. Its second sentence declares that AI will ‘provide key competitive advantages to companies and the European economy’.4 The opening paragraph goes on to state that regulation should ‘preserve the EU’s technological leadership’. Below, we summarize the book’s 17 chapters. In Chapter 1, Shin-yi Peng, Chin-Fu Lin, and Thomas Streinz introduce the three cross-cutting themes informing the discussions on artificial intelligence and international economic law throughout the book: disruption, regulation, and reconfiguration. The chapter provides context for these themes by addressing the (re)emergence of artificial intelligence and the transformation of the global economy, the relevance of international economic law to artificial intelligence regulation, and the reconfiguration of international economic law prompted by advances in artificial intelligence. They conclude by identifying questions not addressed in the book, such as perspectives of developing countries in approaching artificial intelligence and international economic law. Gregory Shaffer (University of California, Irvine) examines dimensions of a global economy built on data as a new form of property, including social challenges, implications for trade law and its negotiating context, and solutions to enhance trade and regulatory efficacy. Eight challenges illuminate what must be addressed in emerging regulatory frameworks. Shaffer proposes one such ‘modest’ framework that ‘foregrounds the importance of building resilience and engaging in problem solving, learning, and adaptation’, which aims to bolster international cooperation and goodwill, as well as equality and democratic values at the national scale. Central to Rolf Weber’s (University of Zurich) argument is that law, as a structural system impacted by changes in the systems with which it interacts, cannot ignore technological developments if policymakers intend to support and stabilize civil society. Likewise, international trade governance, as Weber asserts, should be responsive to advances in technologies such as big data and cloud computing while remaining grounded in substantive legal values, i.e. rights to property and privacy. To balance transparency, trust, and traceability in the international trade regime given these technological developments, Weber argues that an ‘optimal design’ for global trade policy must consider risk assessment, ethical considerations, and other aspects encouraging trust among stakeholders. Regulation should facilitate interoperability between data and technical standards and minimize further fragmentation in the international trade regime. In discussing the political, social, and technical implications of AI technologies, Dan Ciuriak and Vlada Rodionova (both of Ciuriak Consulting Inc.) illustrate the challenges such technologies face in the present international trade regime. Such ‘rites of passage’ include AI’s entry into the global market, clearing hurdles posed by national security, income distributions, and other policy concerns. AI’s final integration is dependent on its success in the preceding steps. The authors predict that the AI integration process will grow more complicated over time due to legal and social questions of the exercise of agency by AI, as well as the outpacing of regulatory efforts by technological development. Aik Hoe Lim (Trade and Environment Division, World Trade Organization (WTO)) identifies Industry 4.0—a shift in global economic processes and production enabled by highly sophisticated and integrated technologies increasingly operating without human involvement—as a driving force in similar shifts in the global trade regime. Key to harnessing its potential is the WTO Agreement on Technical Barriers to Trade (TBT), which has addressed regulatory interventions in trade to the present. The TBT may be further utilized to promote global regulatory coherence and cooperation in light of technological disruption. Crucial to the TBT’s success in accommodating Industry 4.0 is an emphasis on interconnectivity and interoperability to counter discriminatory or divergent global trade regulations. Using connected and autonomous vehicles (CAVs) as examples of ‘disruptive innovation’, Peng addresses the importance and urgency of modernization of the TBT Agreement in response to emerging AI technologies. She identifies and assesses two systematic issues in international trade law that will be shaped by the development of CAVs: goods/services boundaries and public/private sector boundaries. Resulting regulatory frameworks are likely to be shaped by industry-driven standardization, voluntary governance, and the degree of government involvement. Bryan Mercurio (Chinese University of Hong Kong) and Ronald Yu (Chinese University of Hong Kong) pose and preliminarily address the converging legal, data, and policy questions relating to AI-generated inventions and the role of AI in intellectual property regimes generally. To do so, they define AI and identify debates as to its operational contours, evaluate legal norms and precedent impacting AI in the contexts of patents, trade secrets, and copyright and assess issues concerning intellectual property rights in AI system data. The authors conclude that, due to the ‘uneasy fit’ of AI with the international IP regime, AI developers should secure benefits and mitigate risks through contract and the selection of jurisdictions with norms conducive to further AI development. Acknowledging the myriad definitions and applications for ‘digital trade’, Yuka Fukunaga (Waseda University) argues that the character and mechanisms of digital trade disputes will differ greatly from more conventional trade disputes. These differences are evident in the composition of stakeholders, with digital trade encompassing both public and private entities, and the balance between trade and non-trade values, affected by the blurred distinction between goods and services in digital trade. Fukunaga calls for the WTO to consider in its negotiations on trade-related aspects of electronic commerce both substantive digital trade rules and procedural matters, taking into account the new stakeholders in the kinds of disputes raised by digital trade. Fukunaga’s account is compelling, and one hopes that negotiators in the ongoing e-commerce negotiations take note. Focusing his research on the emerging role of data as an ‘essential rent-generating productive asset in the AI economy’, Streinz seeks to advance scholarship on data-related provisions in recent international economic law instruments as well as the regulation of data as a resource. Streinz identifies regulatory interventions implemented by states to derive data’s economic benefits (e.g. open data initiatives), demonstrates the tensions between such interventions and both existing and emerging international trade and investment law commitments, and suggests ways that international economic law could accommodate experimental digital economy policies and counter asymmetric control over data. Hervé evaluates the EU’s unique position, given its history of economic integration and involvement in international trade agreements, as it seeks to accommodate innovations in data governance and AI technologies. Receiving attention are debates regarding data protectionism both generally and in the context of EU legislation, the EU’s role in advancing international digital trade law as compared to competing approaches such as that offered by the United States and the viability of the EU’s approach as a model for similar trade agreements. Whether the future brings convergence or divergence, Hervé asserts, remains unclear as recent events, including the COVID-19 pandemic testing EU states’ capacities for coordination, will impact fragmentation in trade governance. Frederike Zufall (Max Planck Institute for Research on Collective Goods; Waseda Institute for Advanced Study) and Raphael Zingg (Waseda University; ETH Zurich, Center for Law and Economics) critique EU regulations regarding data portability, namely the General Data Protection Regulation, in light of their hypothesis that generated data are more highly valued than raw data. While the regulations categorize data as either personal or non-personal, this distinction is ineffective with regards to developments in data portability. A more effective data portability regime, Zufall and Zingg argue, would distinguish between raw and generated data to create regulations based on the cost-bearing responsibilities of each. Lin examines the implications of regulating automated driving systems (ADSs) using a standardized governance model, which key nations in the ADS value chain have not yet agreed upon. Conflicts arising during the formulation of these rules and standards pose issues to relations within the WTO. Lin analyzes in depth the ethical dimension of ADSs and the subsequent impact on existing and emerging regulatory dilemmas in international trade, including the creation of algorithms to comport with cultural or political norms that circumscribe the permissible choices of ADSs. Neha Mishra (Australian National University College of Law), observing that both governments and private entities have formulated data ethics frameworks confronting increasingly complex data-driven technologies, argues that such frameworks may nonetheless have a trade-restrictive effect. Her research addresses the ability of international trade agreements, especially the General Agreement on Trade in Services (GATS) due to its general exceptions, to provide ‘sufficient policy space’ in the implementation of data ethics policies despite trade-restrictive effects. Mishra envisions a role for WTO panels in applying GATS to accommodate data ethics principles, which may vary by nation, and balancing conflicting domestic and transnational interests. Shang and Du examine the relationship between present WTO law and controversial uses of AI policies, such as restrictions on human rights and contravention of fair competition principles, to evaluate whether WTO law can sufficiently regulate data-sharing policies or justify sanctions against members for the controversial use of AI technologies. The authors propose that WTO law can assist in controlling controversial AI policies by challenging certain data-sharing mechanisms as actionable subsidies or implementing economic sanctions to combat threats to fundamental rights or national security. Following China’s participation in formulating the second Joint Statement Initiative at the 2019 World Economic Forum, Henry Gao (Singapore Management University) analyzes questions related to China’s data regulation policies and its likely positions in future e-commerce negotiations. Gao discerns China’s goals and values relevant to future negotiations by examining its history of e-commerce development, previous statements and submissions to WTO processes, and stated public policy objectives. Noting China’s willingness to comply with certain e-commerce obligations as well as its active role in establishing e-commerce norms, Gao calls for active engagement with China in upcoming negotiations. In light of an emerging ‘Knowledge Revolution’ and a subsequent rivalry between the USA and China in technological advancement, Jane Winn (University of Washington) and Yi-Shyuan Chiang (National Tsing Hua University) explore the possibility of China prevailing as a leader in ‘strategic knowledge competence’. The authors liken the competition between the USA and China to a ‘land rush’, and examine China’s impact on the international trade law regime and its participants through the regime’s emphasis on a pluralist legal culture. Lisa Toohey (Newcastle Law School, University of Newcastle) anticipates in a thought experiment how the Fourth Industrial Revolution, fueled by the impact of data-driven technologies disrupting previously established boundaries, will affect international trade and the workings of the WTO. Three emergent technologies punctuate Toohey’s analysis: artificial intelligence, distributed ledger technologies using blockchain, and the IoT. She ultimately envisions the potential of technological change in allowing the WTO to recalibrate its relationship with both member states and private individuals seeking to utilize international law in trade-related matters. International economic law was largely written in a world where AI was not yet available at commercial scale, so fitting this law to AI can be a complicated task. Much work remains to be done. For example, is it possible to create plurilateral or global frameworks for AI governance? Different national standards both raise costs of global compliance and are susceptible to the impulse to discriminate against foreign AI providers. What is the appropriate international institution to promulgate or enforce any international regulation? Are ISO-type standards5 a useful mechanism to promote high standards in AI? Should AI governance focus on particular applications—finance, insurance, transportation, manufacturing, medical devices, content moderation, toys, etc.—or should we have general-purpose AI regulation? Ciuriak and Rodionova describe the US Food and Drug Administration’s regulatory framework for AI-based medical devices.6 Can mutual recognition systems provide an alternative where international standards do not yet exist? Do we need new forms of transnational dispute resolution to address the new forms of disputes arising from digital trade, as Fukunaga suggests? One possible approach to promoting international standards would be to update the TBT to cover services. As Lim writes, ‘TBT disciplines are unique in the way they put into practice the overarching goal of balancing the right to regulate and the avoidance of unnecessary technical barriers’ (Lim, p. 106). This also requires us to establish that TBT disciplines apply at all. If the regulatory measure is targeted at the artificial intelligence and not at the good itself, it is not clear that the TBT would apply because it is limited to goods.7 Robots in homes and on the streets; computer recommendations about movies and music, books and looks, tweets, and sheets. AI is changing our lives right before our eyes. This book is an important contribution to our understanding of the way that international economic law governs AI. It will certainly be a foundational text for future work. Footnotes 1 Matthew Gault and Joseph Cox, How Amazon’s Astro Robot Tracks Everything You Do, Vice, 28 September 2021, 5:15 pm (quoting an unnamed developer who worked on the robot as saying that ‘Astro … will almost certainly throw itself down a flight of stairs if presented the opportunity’.), https://www.vice.com/en/article/93ypp8/leaked-documents-amazon-astro-surveillance-robot-tracking (visited 25 October 2021). 2 In his chapter, Alan Hervé calls this a kind of ‘European awakening’ (Herve, 194). 3 See Kelly K. Shang and Rachel R. Du, 277. 4 European Commission, Proposal for a Regulation of the European Parliament and of the Council, Laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts, COM (2021), 206. 5 See Dan Ciuriak and Vlada Rodionova, 73–75. 6 Ciuriak and Rodionova, 80–82. 7 Anupam Chander, ‘The Internet of Things: Both Goods and Services’, 18 World Trade Review 9 (2019). For an early call to expand the TBT to cover services, see Anupam Chander, ‘Trade 2.0’, 34 Yale Journal of International Law 281, 324 (2009). © The Author(s) 2021. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Applying International Economic Law to Artificial Intelligence
JF - Journal of International Economic Law
DO - 10.1093/jiel/jgab039
DA - 2021-11-27
UR - https://www.deepdyve.com/lp/oxford-university-press/applying-international-economic-law-to-artificial-intelligence-QfPRISid0T
SP - 804
EP - 809
VL - 24
IS - 4
DP - DeepDyve
ER -