• From battleground to common ground: Bridging Bot shows how AI designed to mediate could nudge toxic online spaces towards healthier discourse – Council on Technology & Social Cohesion.
• Exploring Opportunities for Large Language Models in Public Discourse – Plurality Institute.
• Presentation at ``LLMs in Public Discourse” conference.

• 50-year mortgages, falling real wages, and doing your rideshare due diligence – The Indicator from Planet Money (NPR).

In this paper, I study the Federal Communication Commission’s 2016 Spectrum Incentive Auction, which was developed in collaboration with academic market design economists to reallocate spectrum broadcasting rights from local cable television broadcasters to mobile broadband providers. The paper will proceed in three parts. First, I will analyze the Incentive Auction from the perspective of the performativity of economics (e.g. Mitchell 2007; MacKenzie 2008), focusing especially on the work of ``framing” (Callon 1998)--that is, the work of delineating the economic objects, agents, and decisions that would be ``on stage” and hence relevant in auction transactions. I will show that none of this framing was accomplished ``for free”, but instead required various background legal, political, and material investments to hold in place. Second, I will zoom out and examine the design flexibility of the auction as whole. Of the many ways that the public problem of spectrum reallocation might have been addressed , how and why was the Incentive Auction understood as a necessary, appropriate, and ultimately successful solution? In addressing this question, I will focus especially on the role of ``complexity” in rendering the auction, its particular form, and expert intervention more broadly, as necessary. I will argue that rendering the problem ``complex” is a co-productive move (Jasanoff 2004): it involves both (1) making a fact about the economic reality, and (2) drawing a normative implication about the need for particular sorts of economic expertise in helping to manage it. In the third and final part of the paper, I will consider the auction’s success, exploring how it was rendered successful despite evidence of overflowing (Callon 1998) in the form of strategic manipulation by private equity firms and multi-station owners (Doraszelski et al. 2017).

Resale royalty policies (droit de suite or ``right to follow”) entitle artists to a percentage of secondary market sales of their artwork. These policies have been implemented in many offline art markets (including the UK, EU, and parts of the US) and have been subject to much theoretical debate in economics, law, and policy. At the same time, these policies have recently become popular in digital art markets based on the blockchain, where they are enforced by marketplace platforms rather than governments. Despite their popularity, there is limited high-quality empirical evidence about the consequences of resale royalty policies, partially due to data limitations in offline art markets, such as the inability to observe private sales which represent over half of transactions. In this project I use data from blockchain-based non-fungible token (NFT) art markets to study the effect of resale royalty policies; in particular, I examine the effect of these policies on prices and quantities in the secondary market for NFT assets. To identify causal effects, I take advantage of a recent series of marketplace platform policy changes - driven by competition for buyers - which have substantially reduced effective royalty enforcement in the Ethereum NFT space.