Today, governments and private actors can collect, store, and continuously update vast troves of data. Yet we have barely begun to understand the impact on our democracy of large-scale data collection and the use of such data sets to make decisions that can dramatically impact individual lives and entire communities.

The Data and Democracy Symposium will investigate how technological advances relating to the collection, analysis, and manipulation of data are affecting democratic processes, and it will ask how the law must adapt to ensure the conditions for self-government. Organized by the Institute’s Senior Visiting Research Scholar, Yale Law Professor Amy Kapczynski, and co-sponsored by the Law and Political Economy Project at Yale Law School, the symposium will focus on three areas that are both central to democratic governance and directly affected by advancing technologies and ever-increasing data collection: 1) public opinion formation and access to information; 2) the formation and exercise of public power; and 3) the political economy of data.

Panel links are emailed to participants after they RSVP and are also available below along with other panel information.

Event Materials

Program

Videos

Podcast

Schedule

Thursday October 15

Online
11:45AM – 12:00PM EDT
Welcome and Introductory Remarks

Watch introductory remarks here.
- Jameel Jaffer, Knight First Amendment Institute
- Amy Kapczynski, Yale Law and Political Economy Project; Knight First Amendment Institute
12:00PM – 1:30PM EDT
Theories of Democratic Regulation and Lawmaking in the Big Data Era

How do big data, artificial intelligence, and algorithmic governance influence our ability to govern ourselves democratically? Do new information technologies create new modalities or concentrations of power, and if so what should that mean for how we legislate and regulate? Does our technological landscape challenge conventional paradigms of governance that rely on individualized consent or that see markets as superior because of their ability to reflect and adapt to decentralized information? This panel will address these questions and explore the implications of increasingly data-intensive processes and technologies for how we understand political economy, privacy, and the regulatory state.

Watch panel here.

Panelists
- Yochai Benkler, Harvard Law School
- Kiel Brennan-Marquez, University of Connecticut School of Law
- Julie E. Cohen, Georgetown University Law Center
- Mariano-Florentino Cuéllar, Stanford Law and California Supreme Court
- Aziz Huq, University of Chicago School of Law
- Daniel Susser, Penn State University
Moderator

Amy Kapczynski, Yale Law and Political Economy Project; Knight First Amendment Institute

Privacy, Autonomy, and the Dissolution of Markets

Kiel Brennan-Marquez, University of Connecticut School of Law, and Daniel Susser, Penn State University

Throughout the 20^th century, market capitalism was defended on parallel grounds. Defense one: it allocates resources efficiently. Defense two: it promotes freedom by enabling individuals to exploit their own property. Recently, however, both defenses have begun to unravel—as capitalism has moved into its “platform” phase. Today, the pursuit of allocative efficiency, bolstered by pervasive data surveillance, often undermines individual freedom rather than promoting it. And more fundamentally, the very idea that markets are necessary to achieve allocative efficiency has come under strain. Even supposing, arguendo, that the claim was true in the early 20^th century when Von Mises and Hayek pioneered it, advances in computing have rekindled the old “socialist calculation” debate. And this time around, markets—as information technology—are unlikely to have the upper hand.

All of this, we argue, raises an important set of governance questions regarding the political economy of the future. We focus, specifically, on two: how much should our economic system prioritize freedom—or better yet, autonomy—and to what extent should it rely on markets? The arc of platform capitalism bends, increasingly, toward a system that neither prioritizes autonomy nor relies on markets; and the dominant critical response, exemplified by Shoshana Zuboff’s work, has been to call for a restoration of market capitalism. Longer term, however, it would be more productive to think carefully about how “post-market” economic arrangements might promote autonomy more effectively than markets themselves—and the practical steps necessary to realize that possibility.

How (Not) to Write a Privacy Law

Julie E. Cohen, Georgetown University Law Center

After many years of failed starts, proposed information privacy legislation has begun moving forward in both houses of Congress. As of this writing, the privacy reform docket is crowded, with a number of different proposals jostling for attention and no consensus yet on which deserves to be the front runner. Even so, there is growing agreement on the list of features that a successful bill will need to include. This paper critically assesses those zones of emerging consensus. It argues that many of the shared features of proposed privacy legislation embody fundamentally backward-looking approaches that cannot hope to constrain the activities they attempt to address. They are legal artifacts, designed around the drafting conventions of a prior era and deeply unsuited to contemporary networked realities.

Toward the Democratic Regulation of AI Systems: A Prolegomenon

Mariano-Florentino Cuéllar, Stanford Law and California Supreme Court, and Aziz Huq, University of Chicago School of Law

Cuéllar and Huq explore the challenge of regulating “artificial intelligence” (AI) in democracies. They begin with a definition of what is being regulated. In contrast to the relatively narrow focus on technical details of computational tools in many discussions about governing AI, they suggest that it is more useful to identify “AI systems”—embedding not only particular algorithms but design choices structuring human-computer interaction and the allocation of responsibility over decisions—as the appropriate object of regulation. They make the case that even in constitutional democracies, regulation of these systems should often depend primarily on how these systems embed forward-looking “policies” and on the social consequences of such policies, rather than on expecting clear answers to deontologically flavored questions about whether these systems violate “rights,” such as those to privacy or nondiscrimination. They canvas some of the challenges associated with carefully designed, prudent regulation of AI in democracies. They suggest that as democratic societies contend with the mix of risks and benefits associated with AI systems, candid acknowledgment of the challenges will bring valuable attention to the endogeneity of democratic preferences and to the characteristics of institutions that have an outsized role in shaping how societies evolve.
2:30PM – 4:00PM EDT
Public Will Formation and Big Data

Our changing capacities to capture, analyze, and manipulate data have potentially drastic implications for our ability to exercise self-government and to ensure a truly representative and participatory democracy. How do changes in how we think about and manipulate data change who we understand the public to be, and how it expresses its will? This panel will explore the social and political history of how data has been used for core political objectives like an accurate census count and voter mobilization efforts, how changing technological means to manipulate data are impacting inequality and stratification, and how law and policy should respond.

Watch panel here.

Panelists
- Dan Bouk, Colgate University
- danah boyd, Data & Society
- Bertrall Ross, University of California, Berkeley, School of Law
- Douglas Spencer, University of Connecticut School of Law
- Ethan Zuckerman, University of Massachusetts at Amherst and Knight First Amendment Institute
Moderator

Katy Glenn Bass, Knight First Amendment Institute

Democracy’s Data Infrastructure: The Entanglement of Politics and Science

Dan Bouk, Colgate University, and danah boyd, Data & Society

Bouk and boyd examine the technopolitics of the U.S. census, showing that despite their efforts, Census Bureau scientists and their peers in the academy have always ended up working in the political (not to be confused with partisan) sphere. Bouk and boyd use three case studies to illustrate this point: (1) the formalization of congressional apportionment and the rise of professional statistics and tools for dealing with aggregates; (2) the relationship between the problem of the “undercount” and the probabilistic revolution of the mid-20th century, and the combined influences of microeconomics and data science on the census; (3) the difficulty of avoiding the disclosure of confidential information and the current debate over Census Bureau use of differential privacy. This series of technopolitical controversies, they argue, demonstrates that the legitimacy of data infrastructure depends as much on social, political, and historical factors as on the quality and accuracy of the data.

Voter Data and Democratic Inequality

Bertrall Ross, University of California, Berkeley, School of Law, and Douglas Spencer, University of Connecticut School of Law

The mobilization activities of parties and candidates prior to elections can be a critical vehicle for encouraging participation. Thanks to recent advances campaigns have shifted mobilization efforts toward microtargeting strategies based on individual and household level data rather than neighborhood data. One of the most valuable pieces of information is whether someone voted in prior elections. Ross and Spencer trace the history of voter history data and highlight the role of federal and state law in facilitating campaign access to it. They then examine the consequences. Drawing on a novel dataset of voter history laws in every state, Ross and Spencer find that after states change their laws to provide voter history to campaigns, these campaigns are far less likely to contact the poor. Because the poor are less likely to have voted in prior elections, they are less likely to appear in the mobilization models employed by data-savvy campaigns, exacerbating the rich-poor turnout gap. To address this, Ross and Spencer advance two proposals. First, they argue that states should prohibit the collection and distribution of voting history data. However, because campaigns might respond by using proxies for individuals’ voting histories, they suggest an alternative. States could continue to collect and distribute voting history data, but create financial incentives for campaigns to use the data to expand their mobilization strategies to a more representative target population.
Friday October 16

Online
12:00PM – 1:30PM EDT
Data Publicity and Public Control

Government actors are today allocating Medicaid resources, evaluating employees, conducting surveillance and criminal investigations, and making bail decisions using algorithmic and data-intensive approaches. These approaches often involve partnerships with private actors who control the development of the technology being used and the access to relevant data. This panel will contemplate whether the resulting systems can be reconciled with commitments to due process and nondiscrimination, what challenges they raise for accountability, and how courts, legislatures, and agencies should address those challenges to protect democratic prerogatives.

Watch panel here.

Panelists
- Margot Kaminski, University of Colorado Law and Silicon Flatirons
- Martin Murillo, Institute of Electrical and Electronics Engineers (IEEE)
- Frank Pasquale, Brooklyn Law School
- Wendy Wagner, University of Texas School of Law
- Rebecca Wexler, University of California, Berkeley, School of Law
Moderator

Paul Ohm, Georgetown University Law Center

Licensing Big Data Analytics in an Era of Invasive and Contested AI

Frank Pasquale, Brooklyn Law School

Pasquale argues that the use of big data analytics necessitates regulatory action, specifically arguing in favor of a licensing regime. He starts by highlighting the opportunities and threats posed by AI, exploring examples in the health industry and facial recognition. Pasquale lays out the many weaknesses of the current, consent-focused regulatory model. Instead, he argues in favor of a governance-driven approach that would explicitly bar certain collection, analysis, and use of data, while promoting other, more socially useful analytics. Many civil libertarians have rallied behind restrictions on facial recognition; this logic may also apply to limitations on other forms of recognition by data (for example, diabetic recognition by employers). Using Senator Sherrod Brown’s DATA Act as a case study, he provides theoretical rigor and practical specificity for the proposal. Pasquale ends by discussing the threat free-expression fundamentalism poses to licensing regimes in particular and data regulation in general. He argues that in many instances data processing is less a form of communication than an exertion of power. Free expression law must recognize this critical distinction in order to promote democratic control over data policy.

Is the Administrative State Ready for Big Data? Exploring the Accountability Challenges that Big Data Tools Present for Environmental and Public Health Regulation

Wendy Wagner, University of Texas School of Law, and Martin Murillo, Institute of Electrical and Electronics Engineers (IEEE)

Murillo and Wagner look at the administrative state’s growing use of complex statistical models and the challenges the trend poses for accountable administrative governance. They document how agencies’ use of big data can obscure critical framing decisions underlying policies, hide subjectivity in the design and development of models, and undermine scientific integrity. Legal process requirements should in theory counteract these tendencies to sideline public deliberation and oversight. But in practice, they argue, administrative process tends to perversely reward agencies for using models that are inaccessible to the public due to the threat of judicial review and protracted comment processes. To address these challenges, they propose a formal, interdisciplinary review process that requires agency staff to comprehensibly explain—using best practices—the framing, algorithm choices, and procedures used to ensure the integrity of their analyses. They also suggest using rewards such as increased judicial deference for accessible explanations to incentivize quality and transparent models.

The Law and Political Economy of Privacy Asymmetries

Rebecca Wexler, University of California, Berkeley, School of Law

Wexler focuses on how law enforcement’s capacity to capture and analyze data is far surpassing that of the accused. Recent developments in cross-border access to data aggravate this asymmetry by creating new access mechanisms available to law enforcement alone. Meanwhile, privacy justifications can be used superficially to withhold relevant evidence from the accused. Wexler’s central claim is that exceptions to privacy protections that enable law enforcement to access data across borders should also apply to defense investigators. She argues that the asymmetrical expansion of law enforcement’s investigative power is an unreasoned, likely unintentional, and largely unjustified side effect of the political economy of bilateral negotiations. Wexler ends with two constitutional arguments. First, principles of due process restrict the government from enacting privacy regulations that foreclose defendants’ access to exculpatory data while maintaining law enforcement’s access to inculpatory data. Second, such asymmetrical access regimes should be presumed unconstitutional unless the government provides nonarbitrary and proportionate justifications for the asymmetry.
1:40PM – 2:30PM EDT
Keynote

Watch keynote here.
- Julia Angwin, The Markup
2:30PM – 4:00PM EDT
Engineering Access to Data and Algorithms

Concerns over the spread of disinformation, misinformation, and hate online have never been higher. Similarly, the ongoing adoption of AI and machine learning tools by local, state, and federal government has generated increasing pressure for public access to information about those tools and how they are being used. Journalists, researchers, and activists say they do not have sufficient ability to access the data and information needed to hold government and powerful companies to account. But these calls for more transparency and access often collide with objections by private companies that they cannot release more data to researchers and/or the public due to privacy concerns or because the information constitutes “trade secrets.” What approaches might help secure better access to privately held data for researchers and the public in practice? This panel will explore this question, and what it implies about the limits of traditional understandings of transparency and accountability.

Watch panel here.

Panelists
- Hannah Bloch-Wehba, Texas A&M University School of Law
- John Bowers, Harvard Berkman Klein Center
- evelyn douek, Harvard Law School and Harvard Berkman Klein Center
- Elaine Sedenberg, Harvard Berkman Klein Center and Facebook
- Mathias Vermeulen, AWO
- Jonathan Zittrain, Harvard University
Moderator

Alex Abdo, Knight First Amendment Institute

Transparency’s AI Problem

Hannah Bloch-Wehba, Texas A&M University School of Law

Bloch-Wehba argues that AI reflects and amplifies systemic weaknesses in the transparency regime, including privatization, secrecy, private sector cooptation, and reactive disclosure. This highlights the urgent need to reorient transparency and accountability law toward meaningful public engagement in ongoing oversight. She says this shift requires rethinking FOIA’s core commitment to public disclosure of agency records, exploring instead alternative ways to empower the public and shed light on decision making. Bloch-Wehba uses criminal law enforcement as a case study for these issues, illustrating the difficulties activists face in obtaining information about AI and the limits of a technocratic approach to transparency. Finally, she lays out two proposals for addressing these issues. First, she suggests that states and municipalities reform procurement laws to consider factors beyond price such as the vendor’s commitment to openness. Second, she proposes new forms of proactive disclosure that foster openness in AI, similar to efforts to promote open science.

Platform Accountability Through Digital “Poison” Cabinets

John Bowers, Harvard Berkman Klein Center; Elaine Sedenberg, Harvard Berkman Klein Center and Facebook; and Jonathan Zittrain, Harvard University

Bowers, Sedenberg, and Zittrain draw on the history of a centuries-old German archival institution, the Giftschrank (literally “poison cabinet”), to explore potential means of surmounting barriers to data sharing faced by contemporary internet platforms—particularly those engaged in content moderation. Dating back to the 16th century, Giftschränke were restricted sections of libraries stocked with texts perceived as dangerous to the public, often accessible only to properly credentialed researchers seeking to understand or countervail them. Today’s platforms, reasonably wary of the prospect of developing a record of the very content they deem unacceptable, could—as an alternative to a status quo defined largely by deletion—build their own moderation-focused Giftschränke, preserving detailed records of what they take down and why. The platforms might embargo the release of such archives, before ultimately making them available to academic researchers, enabling better-informed research into content ecosystems over the long term. Drawing on the lessons, innovations, and public spirit of the archival sciences, the authors examine the risks, upsides, and constraints of such an approach.

The Keys to the Kingdom: Overcoming GDPR Concerns to Unlock Access to Platform Data for Independent Researchers

Mathias Vermeulen, AWO

Independent researchers have requested access to a wide variety of platform data in order to scrutinize the actors and methods responsible for the spreading of disinformation, as well as to assess the effectiveness and impact of the self-regulatory measures platforms have taken to counter disinformation. Over the past two years the European Union has increasingly supported these requests, including by facilitating the adoption of the EU Code of Practice on Disinformation, in which platforms committed themselves to provide such access through a variety of mechanisms. However, both European decision makers and researchers have consistently criticized the insufficient responses by some platforms - in particular in handing over proprietary data. On a number of occasions platforms have invoked their obligations under the EU’s General Data Protection Regulation (GDPR) as rationale for failing to make certain data sets available. This paper aims to assess those claims in order to clarify to which extent, if any, GDPR concerns prevent the disclosure of (sensitive) personal data by platforms to independent researchers. The paper will argue that the GDPR currently allows the sharing of personal data for such purposes in a number of ways, but recommends the adoption of a binding legal obligation for platforms to hand over data and a Code of Conduct on Access To Platform Data under Article 40 of the GDPR as necessary measures to provide more legal certainty about the specific roles and responsibilities of both platforms and independent researchers.