Research

Broadly, my research lies at the intersection of value theory and social science (particularly economics and game theory). In addition to a number of other smaller projects, I have two ongoing research programs.

Toward an "ecological" doctrine

My immediate agenda consists in converting my dissertation into a book manuscript.

The first two chapters of the dissertation are an examination of a variety of serious moral problems---such as slavery, the appropriation of indigenous lands, rural flight, demographic pressure, poverty, underdevelopment, and the general lawlessness of AI technologies---through the lens of two broad sorts of overall philosophical view. One sort of view, which I simplify under the heading of the "economic doctrine," represents Milton Friedman, Murray Rothbard, and Robert Nozick. The other, which I call the "political doctrine," represents John Rawls and Philip Pettit.

In a current project, I sketch a view I call the "ecological doctrine," which upholds a naturalized form of transnational republicanism, grounded in substantive accounts of law, freedom, and justice. I argue that, in comparison to the other two doctrines, my ecological doctrine offers a more satisfying and conceptually unified solution to some of our most serious moral and political problems. The methodology of the ecological doctrine, so to speak, has a long pedigree: back in 1956, Herbert Simon offered the formal framework for a realistic model of human world-simulations. However, computer science and technology have only recently developed to such an extent that it is computationally feasible to realize his vision. We can now use ordinary desktop machines to model complex emergent social phenomena at global scale, in terms of interactions among simulated agents whose beliefs, preferences, affective states, constraints, and value functions resemble ours (see Josh Epstein's Agent_Zero: Toward Neurocognitive Foundations for Generative Social Science), in a realistic simulated natural environment (see Joseph Suarez's ``Neural MMO'' platform).

Currently most of the interest among the developers of these tools is in computational social science, but my hypothesis is that the world-simulation framework could also be used as a tool for exploring deeper issues in value theory. Once we are capable of producing models that embody the emergence of complex compositional languages, we will be able to deploy the models to explain the natural emergence of legal and moral norms, perhaps in the form of a statistical demonstration of H.L.A. Hart's ``minimum content of natural law." My eventual plan is to use these world-simulations to map the space of empirically feasible social and institutional forms.

AI ethics

One immediate project in information ethics involves completing the manuscript for the book I am co-authoring with Clinton Castro and Alan Rubel. This book, which is titled Algorithms and Autonomy: the Ethics of Automated Decision Systems (accepted for publication in 2021 with Cambridge University Press), draws in part from our published and unpublished articles.

In the first third of the book, we discuss what agency is and how we can secure it, in light of the pervasiveness of automated decisions systems and digital platforms. Our discussion begins with a discussion of the nature of autonomy, drawing largely from "Algorithms, Bias, and the Importance of Agency" (in Social Theory and Practice). We discuss three ways automated decision systems conflict with agency. They govern people's behavior in ways that they could not reasonably abide, conflict with people's agency by denying them information to which they are entitled, and fail to respect the boundaries between persons.

In the second third of the book, we turn to the topic of freedom and how to protect it. In "Information Technologies and Freedom as Undiminished Agency," we offer a theory of freedom that is specified in terms of three sorts of human limitation, and in ``Epistemic Paternalism Online" (in the volume Epistemic Paternalism Reconsidered) we discuss to the policy agenda implied by this theory of freedom.

In the final third of the book, we turn to the responsibilities of agents. In "Agency Laundering and Information Technologies" (in Ethical Theory and Moral Practice), we examine what we call ``agency laundering"; the distancing of oneself from morally suspect actions by deflecting one's moral responsibility for them. We employ this concept to clarify the issues in a number of recent public cases involving platforms and other systems. Then, in "Democratic Obligations and Technological Threats to Legitimacy," we turn to the realm of politics and the question of civic obligation: we examine our responsibilities to support responsible democratic governance in the age of predictive policing and Cambridge Analytica.

In addition to the book project (and the additional articles that we are currently developing), another ongoing project in AI ethics involves my philosophical work about noxious economic markets. In "The Moral Limits of the Market: The Case of Consumer Scoring Data" (in Ethics and Information Technology) and "Is the Attention Economy Noxious?" (in Philosophers' Imprint), Clinton and I offer a justification for the regulation of algorithms, on the basis of the noxious character of the market for personal data. This market, which is primarily driven by a business model in which people pay for online services by their very consumption of it, has a number of troubling ethical and social implications.

My Papers