*The Research Roundup is a semi-regular list of outside research we have found interesting and think is worth sharing. The views and conclusions of the papers’ authors do not necessarily reflect the opinions of anyone affiliated with TPI.*
This month’s articles focus on two sub-fields of technology policy: the social and economic implications of artificial intelligence (AI), and questions of intellectual property (IP) in the digital space. In Demographics and Automation, Daron Acemoglu and Pascual Restrepo track parallel trends in industrial automation and demographic changes. Workers over 36, they suggest, are being replaced not by younger workers, but by robots. Additionally, countries with the greatest numbers of older workers have invested most heavily in industrial automation systems. They find that this pattern varies across industries based on the applicability of automation, so older workers in fields like nursing and education are less likely to be replaced than are workers in manufacturing or finance positions. Given the size of the U.S. population at or nearing retirement age, this relationship could have important implications for the future of work in America. Click through to read more about this and other papers exploring AI and IP.
Descriptions of papers below are edited abstracts from authors
Artificial Intelligence: Challenges, Benefits, and Questions
The authors argue theoretically and document empirically that aging leads to greater (industrial) automation, and in particular, to more intensive use and development of robots. Using US data, we document that robots substitute for middle-aged workers (those between the ages of 36 and 55). They then show that demographic change—corresponding to an increasing ratio of older to middle-aged workers—is associated with greater adoption of robots and other automation technologies across countries and with more robotics-related activities across US commuting zones. They also provide evidence of rapid development of automation technologies in countries undergoing greater demographic change. Their directed technological change model further predicts that the induced adoption of automation technology should be more pronounced in industries that rely more on middle-aged workers and those that present greater opportunities for automation. Both of these predictions receive support from country-industry variation in the adoption of robots. Their model also implies that the productivity implications of aging are ambiguous when technology responds to demographic change, but we should expect productivity to increase and labor share to decline relatively in industries that are most amenable to automation, and this is indeed the pattern we find in the data.
Personal mobility is facing three major innovations that have disruptive potential: electrification, shared mobility, and automation. The author presents each innovation individually and focuses on their role(s) in disrupting the auto industry, the transport system, and energy system. The largest disruptive potential lies in the combination of the three innovations, i.e., in the shared autonomous electric vehicles (SAEV). While shared mobility per se might not have the potential to truly disrupt the transport system it is necessary to steer electrification and automation in a more sustainable direction. Technology and innovation alone will not be sufficient to create a new sustainable transportation system. Regulations will also be necessary.
Edmond Awad, Sydney Levine, Max Kleiman-Weiner, Sohan Dsouza, Joshua B. Tenenbaum, Azim Shariff, Jean-François Bonnefon, and Iyad Rahwan
When a semi-autonomous car crashes and harms someone, how are blame and causal responsibility distributed across the human and machine drivers? In this article, we consider cases in which a pedestrian was hit and killed by a car being operated under shared control of a primary and a secondary driver. We find that when only one driver makes an error, that driver receives the blame and is considered causally responsible for the harm, regardless of whether that driver is a machine or a human. However, when both drivers make errors in cases of shared control between a human and a machine, the blame and responsibility attributed to the machine are reduced. This finding portends a public under-reaction to the malfunctioning AI components of semi-autonomous cars and therefore has a direct policy implication: a bottom-up regulatory scheme (which operates through tort law that is adjudicated through the jury system) could fail to properly regulate the safety of shared-control vehicles; instead, a top-down scheme (enacted through federal laws) may be called for.
Recent rapid advances in Artificial Intelligence (AI) and Machine Learning have raised many questions about the regulatory and governance mechanisms for autonomous machines. Many commentators, scholars, and policy-makers now call for ensuring that algorithms governing our lives are transparent, fair, and accountable. Here, the author proposes a conceptual framework for the regulation of AI and algorithmic systems. He argues that we need tools to program, debug and maintain an algorithmic social contract, a pact between various human stakeholders, mediated by machines. To achieve this, we can adapt the concept of human-in-the-loop (HITL) from the fields of modeling and simulation and interactive machine learning. In particular, he proposes an agenda I call society-in-the-loop (SITL), which combines the HITL control paradigm with mechanisms for negotiating the values of various stakeholders affected by AI systems, and monitoring compliance with the agreement. In short, ‘SITL = HITL + Social Contract.’
IP in the Digital Space
Bruno Jullien and Yassine Lefouili
This paper discusses the effects of horizontal mergers on innovation. The authors rely on the existing academic literature and our own research work to present the various positive and negative effects of mergers on innovation. Their analysis shows that the overall impact of a merger on innovation may be either positive or negative and sheds light on the circumstances under which each of these scenarios is likely to arise. They derive a number of policy implications regarding the way innovation effects should be handled by competition authorities in merger control and highlight the differences with the analysis of price effects.
Luca Aguzzoni, Benno Buehler, Luca Di Martile, Ron Kemp, Anton Schwarz
Recently there has been an increased attention towards the ex-post evaluation of competition policy enforcement decisions and in particular merger decisions. In this paper, the authors study the effects of two mobile telecommunication mergers on prices. They apply a standard difference-in-differences approach, which is widely used in the literature on ex-post evaluation of mergers. For the Austrian T-Mobile/tele.ring merger, they conclude that after the acquisition (for which remedies were imposed) prices in Austria did not increase relative to the considered control countries. For the Dutch T-Mobile/Orange merger, they observe an increase in the mobile tariff prices in the Netherlands in the analyzed period, relative to the control countries.
This essay investigates the development of peculiar models in patents: “Open patent” licenses, patent pledges, defensive patent strategies and the like. Although using patents in a nonconventional way, these models do not dispute the very existence of patents, but rather the way patents are being used. These models are not entirely new—some have been around for over a decade—but seem to find an echo with the growing criticisms of the “broken patent system”. These models might be able to go beyond the pro-patent/against patent debate, by reconciling exclusive/inclusive mechanisms and eventually reveal the strong plasticity of patents, offering patents a new lease on life.
P. Jean-Jacques Herings, Ronald Peeters, and Michael S.Yang
This paper uses a dynamic stochastic model to solve for the optimal pricing policy of music recordings in the presence of P2P file-sharing networks eroding their sales. We employ a policy iteration algorithm on a discretized state space to numerically compute the optimal pricing policy. The realistically calibrated model reflects the real-world figures we observe and provides estimates of the optimal pricing policy as well as comparative statics figures. The pricing policy is such that, for a given P2P network size, prices are increasing in the number of buyers of the product and, for a given number of buyers of the product, prices are non-monotonic in the P2P network size. Surprisingly, in the presence of P2P networks, increases in production costs and decreases in the valuation of the product increase the consumer and total surplus. A higher valuation of the product leads to a lower steady state price. Increased switching costs have a negative effect on prices and profits, so the long-term incentive to attract new consumers dominates the short-term incentive to harvest loyal consumers. The full enforcement of intellectual property rights has adverse effects on both consumer surplus and total welfare.