What Are the Implications of Emerging Technologies in AI-Driven Robotics and Automation for Globalization?

By Kenneth Anderson
Sunday, April 9, 2017, 2:51 PM

Last month I attended an outstanding workshop at the University of Pennsylvania’s newly-established Perry World House on a topic that (as far as I can tell) has not received the attention it should: the intersection of emerging technologies with globalization.  As it happens, I work in both these areas—but when I received the workshop invitation to join a panel on the interactions of emerging technologies in AI-enabled robotics and automation with processes of globalization, I realized I had never given much thought to how these two might interact with each other, today and into the future.  

While there is a rapidly growing body of research—and lots of controversy—over the likely effects of robotics and automation on jobs and employment, generally speaking the jobs-automation arguments center on effects within a particular country (the EU being a special case) and its national economy.  There is a lot of cross-national comparison, and a fair amount of discussion on cross-border labor, jobs, outsourcing, and related flows, but these discussions tend to be cross-national, rather than “global,” meaning that the locus of impacts is still national economies.

Although the “global economy” or “economic globalization” is often taken to mean simply the aggregation of these national economies, in globalization studies a distinction is also (sometimes) made between “global” and “international,” as terms of art.  On the one hand, the genuinely “global” conveys the concept of universality, “above” the national level; it is importantly a conception of governance understood to reside at a level above nation-states, and which is not dependent on nation-states or even aggregations of nation-states for its legitimacy and authority.  On the other hand, this meaning of “global” is sometimes distinguished from the merely “international,” which would include national economies, the aggregation of national economies, and governance that is shared by nation-states, as opposed to yielding to institutions of global governance that are genuinely “above” sovereign states and even above the aggregation of nation-states.  This term-of-art sense of “global,” moreover, is often used not merely in a descriptive way, but in a prescriptive way, usually in service to a normative ideal of liberal internationalism as the form of global governance that, it is believed, ought over time to take hold.

It would be something of an understatement to say I've long been a critic of liberal internationalism as the ideal form of international order, law, and institutions—and equally a skeptic about the degree to which liberal internationalism actually has long-term purchase and grip as something toward which the global order is, in point of fact, gradually evolving.  But in relation to the question of advancing technology, economic globalization, and political globalization (aka global governance), it seems beyond question that the technologies that emerged and came of age in the 1990s—communications technologies, the creation and spread of cyber and internet technologies worldwide, and also software of all kinds—were important, indeed key, drivers of economic and political globalization, as they also emerged in the 1990s.

The question that seemed to me most important to consider on the Perry World House workshop panel in which I participated was whether today's emerging technologies of AI-enabled robotics and automation would tend to have the same effects—driving globalization forward in both economic and political senses.  Much of our panel’s discussion revolved around the immensely important question of the effects of these technologies on employment within national economies, and in the aggregation of national economies.  But one of the striking features of this workshop was that it asked questions about emerging technologies in relation to globalization itself, not simply about effects at the national level.

My short, tentative answer is that these technologies of robotics and automation are not necessarily drivers of globalization. Indeed, they might strongly push economic and political globalization away from genuinely “global” structures and toward national ones—cross-border and “international” in their effects, rather than “globalizing” (as terms of art described above).  Why?  Fundamentally because the specific technologies at issue (and there are many other types of emerging technologies to consider, to be sure), robotics and automation, are not communications technologies or purely digitized, software technologies. Though it’s true that “robotics” and “automation” are often used to refer to purely software or cyber agents, the robotics and automation technologies at issue here are physical machines doing physical tasks in the physical world in which they are physically located.

Often these machines are used to produce things, but increasingly they are designed to perform services (e.g., the automated hamburger cooking/flipping machine for use in a fast food restaurant).  These new technologies of robotics and automation, because they are not about cheap (and more or less unstoppable) communications and digitization, do not in and of themselves drive cross-border interactions and transactions.  They might lead to more trade across borders in goods and services, but at bottom they are about place and location and physical things or physical services. These emerging technologies are thus, by definition, about “embodiment” of the machine and its production, not about digitization and the (naively, to be sure, imagined) placeless-ness of cyberspace.  (Cue John Perry Barlowetc.)

In that regard, the effect on the world at large of these robotics and automation technologies tends toward the “international” rather than “global.” This suggests that physical robotics and automation are fundamentally about the “particular” rather than the “universal” (using terms, once again, from globalization theory). Universal communications, at least in its idealized 1990s form, by contrast, tends to correlate with ideals of both liberal internationalist global governance and a tech vision of a universal cyber-connected world, built upon cost-less communications across all distances and parties, and relatively indifferent to (but, in any case, unstoppable by) national jurisdictions.   

My personal view, for what it’s worth, is that it's too soon to tell what large-scale effects robotics and automation, enabled by increasingly powerful AI, will have on jobs and employment, whether in the U.S. economy or elsewhere.  Most economists, so far as I can tell, remain skeptical of the view (which tends to dominate the general-audience writing on this subject) that, even over the long run, “this time is different” and so these technologies will bring about long run or permanently lowered employment.  I tend to the skeptical view as well—but for an additional reason.  I'm (also) skeptical about how quickly or the extent to which robotic machines will be able to fully replace human beings in many different settings. While it's true there's often a "glass half-full/glass half-empty" feel to these debates about robotics and its capabilities now and "sometime" in the future, the field (and AI as well) also often seems to me caught in the grip of “technology enchantment."  Robotics and AI have the strange capacity to lead otherwise sensible people to move, with no warning or justification, from talking hard-headedly about what is likely to occur in the short-to-medium term, on technological paths that are already visible and in place, even if not completely developed—to suddenly embracing science fiction speculation and visions of a world that is, at best, a hundred years in the future.

However, even on the hard-headed, short-to-medium term view of robotics and automation technology, if there are significant job losses on account of robotics and automation, then beyond the general sense that physical machines emphasize the “particular” as a matter of place and location rather than the “universal,” there is perhaps another way in which these technologies do not press toward global political institutions, but instead toward national ones.  Displaced workers are not likely to look, or have any reason to look, to global political institutions of governance to help them with the loss of jobs and income; they will look to their national governments.  It’s an open question whether national governments are able to do much about this in a world that is far more globalized as to free movement of capital, goods, and increasingly labor and people, with their own disruptions of labor markets, than it is with respect to governance.

But displaced workers in any case would look not to global institutions that, in conformity with universal cosmopolitan ideals, fail to favor these workers over anyone else any place else on earth, but instead quite obviously to national governments that they expect to represent their interests and put them above those of non-members of their political community.  The interests of displaced workers might conflict with the interests of others in that society receiving cheaper products or services, or receiving the profits from such sold abroad, but the point is that the dispute takes place within a national society and its institutions. The result of such social pressures is not likely a press toward global institutions, global governance, and universality from above, but instead pressures on national states to assert themselves within the international system on behalf of the political community of which they are a part and a representative—a system of sovereigns that, though acting multilaterally in many things, does not see itself as being or as progressing toward some genuinely global system.

Something like this, after all, underlies at least part of the trend toward national populism in many different societies today. Elites in many places, particularly in developed Western countries, might still gravitate toward universal cosmopolitanism as the most admirable social and political ideal, but those they govern appear attached to social and political understandings of community that are far more particular than that.  If robotics and automation turn out to be major economic players down the road, and especially if they result in significant and long term disruptions of national labor markets, there might indeed be implications for actors in the international or global spheres, and for both economic and political processes of globalization.

(Perry World House is an exciting new endeavor at the University of Pennsylvania, with a lovely brand-new building, and a cross disciplinary orientation in addressing many topics of globalization and global affairs.  Its director and associate director are well-known to many Lawfare readers—Penn law professor Bill Burke-White and Penn political science professor Michael Horowitz—and I’d like to thank them for inviting me to participate in a terrific workshop in Perry World House’s early days.  There were other panels on the program that were also outstanding, on topics ranging from AI and autonomous weapons to cybersecurity and cyberwarfare.  Likewise the event attracted a wonderfully cross-disciplinary group of participants, from fields of engineering to national security policy, and including many old friends of mine. But the aspect of this workshop that most intrigued me was asking how these emerging technologies potentially interact with existing theories and understandings of globalization.)