About

I am an Associate Professor in the School of Public Policy at the Georgia Institute of Technology, the director of the Ethics, Technology, and Human Interaction Center, and a Faculty Affiliate in the Georgia Tech Center for Machine Learning. I lead a team to facilitate the early identification and management of potential ethical and societal consequences of AI-enabled manufacturing systems. My research explores the relationships between values in science, technology, and medicine; the epistemic implications of the social organization of research; and ethics and policy, particularly in the fields of artificial intelligence and machine learning.

Research topics

• Political Science
• Sociology
• Computer Science
• Engineering
• Law
• Public relations
• Management science
• Public administration
• Epistemology
• Management
• Philosophy
• Engineering ethics
• Economics
• Knowledge management
• Criminology

Selected publications

• Artificial Intelligence

  2026-03-10

  book-chapter1st authorCorresponding

  Artificial intelligence (AI) systems are not, and cannot be, value-free. They reflect human values. Given this, and given the potential of AI to impact and disrupt society in significant ways, it is important that the values that influence AI systems are critically scrutinized. This chapter has two primary aims. The first is to examine some important respects in which AI systems are value-laden. More specifically, the chapter examines the influence of values at four stages in the AI lifecycle: (1) problem framing and operationalization; (2) data; (3) modeling and validation; and (4) deployment. This examination paves the way for a second task, which is to address the question of how AI should be governed—including whether there should be moral or political “guardrails” on AI. There is much disagreement surrounding this question, and this chapter will not attempt to answer it definitively. Rather, it will discuss three recent approaches to AI governance, which shed light on ways that policymakers are attempting to address practical questions about the management of values in AI. This discussion will illustrate how controversial and important these questions are. Readers may be interested in these Handbook chapters as well: Bennett Holman and T. Y. Branch, “Reflecting on Responses to the New Demarcation Problem”; Stephen John, “Transparency in Science”; Alexander Tolbert, “Algorithmic Abolitionism and The Racial Algorithm.”

  PublisherDOI

• Enabling Federated Access to Administrative Data and Beyond

  International Journal for Population Data Science · 2025-08-28 · 1 citations

  articleOpen access

  ObjectivesEnable federated access to datasets across multiple Trusted Research Environments (TREs) while adhering to the Five Safes framework and international standards. This approach enhances global interoperability, prioritizing flexibility, accessibility, and transparency to advance secure, collaborative research. MethodThrough DARE UK funding, we developed TRE-FX, a rapid prototype leveraging GA4GH Task Execution Service (TES) for federated analysis. Originally designed for genomics, we extended TES to diverse data types and environments. TRE-FX supports complex workflows while integrating SQL and GraphQL to simplify adoption. Wizard-driven interfaces and API access further enhance usability, enabling seamless execution of federated queries across secure environments. ResultsTRE-FX establishes a hub-and-spoke configuration for secure data access. Safe Settings, Projects, and People ensure controlled access to Safe Data, with results passing through an airlock for authorized release. Open-sourced and backed by further investment, TRE-FX will evolve into a national, cross-domain solution. Transparency is a core principle, with public visibility of projects, participants, and queries fostering trust. This commitment will expand in the next development phase. ConclusionTRE-FX delivers an end-to-end federated analysis solution within nine months, spanning three institutions. By integrating accessible tools like SQL, we support diverse research needs. Future phases will enhance reproducibility with RO-Crates and explore FAIR federated analytics. This initiative drives national dialogue, shaping the future of federated research in the UK.

  PublisherOA PDFDOI

• Challenges and Opportunities in Advancing Equity in Georgia Manufacturing

  2024-10-04

  preprintSenior author

  Recent economic development policy in the United States has, in a break from previous decades, identified equity as an explicit goal. However, little is known about what novel challenges and opportunities face economic development programs and practitioners in attempting to advance equity. This paper reports results from a workshop engaging 48 managers and staff working on components of the Georgia Artificial Intelligence Manufacturing Project (Georgia AIM), a $65 million equity-focused regional economic development project funded by the U.S. Economic Development Administration and the Georgia state government. Participants discussed their understandings of economic participation and equity, potential obstacles and opportunities in achieving equity, and approaches to assessing equity outcomes. Results reveal diversity in understandings of equity in the manufacturing economy. Participants agree that widespread participation in manufacturing work partly constitutes manufacturing equity, but they exhibit more divergence about social, psychological, or environmental aspects of participation and equity. Meanwhile, participants identified a wide variety of both internal and external factors which will affect whether Georgia AIM achieves its equity goals. This report illustrates several tensions which equity-focused economic development projects will have to face—between promoting all or only some sorts of manufacturing jobs; between promoting benefits from manufacturing and reducing harms from manufacturing; and between achieving and assessing impact. Participants’ grounded perspectives on negotiating these tensions can provide guidance for future economic development projects.

  PublisherDOI

• Beyond Algorithmic Fairness: A Guide to Develop and Deploy Ethical AI-Enabled Decision-Support Tools

  arXiv (Cornell University) · 2024-09-17

  preprintOpen access

  The integration of artificial intelligence (AI) and optimization hold substantial promise for improving the efficiency, reliability, and resilience of engineered systems. Due to the networked nature of many engineered systems, ethically deploying methodologies at this intersection poses challenges that are distinct from other AI settings, thus motivating the development of ethical guidelines tailored to AI-enabled optimization. This paper highlights the need to go beyond fairness-driven algorithms to systematically address ethical decisions spanning the stages of modeling, data curation, results analysis, and implementation of optimization-based decision support tools. Accordingly, this paper identifies ethical considerations required when deploying algorithms at the intersection of AI and optimization via case studies in power systems as well as supply chain and logistics. Rather than providing a prescriptive set of rules, this paper aims to foster reflection and awareness among researchers and encourage consideration of ethical implications at every step of the decision-making process.

  PublisherOA PDFDOI

• Report on the Conference on Ethical and Responsible Design in the National AI Institutes: A Summary of Challenges

  arXiv (Cornell University) · 2024-07-18

  preprintOpen accessSenior author

  In May 2023, the Georgia Tech Ethics, Technology, and Human Interaction Center organized the Conference on Ethical and Responsible Design in the National AI Institutes. Representatives from the National AI Research Institutes that had been established as of January 2023 were invited to attend; researchers representing 14 Institutes attended and participated. The conference focused on three questions: What are the main challenges that the National AI Institutes are facing with regard to the responsible design of AI systems? What are promising lines of inquiry to address these challenges? What are possible points of collaboration? Over the course of the conference, a revised version of the first question became a focal point: What are the challenges that the Institutes face in identifying ethical and responsible design practices and in implementing them in the AI development process? This document summarizes the challenges that representatives from the Institutes in attendance highlighted.

  PublisherOA PDFDOI

• How Can We Know if You are Serious? Ethics Washing, Symbolic Ethics Offices, and the Responsible Design of AI Systems

  Canadian Journal of Philosophy · 2024-05-01 · 7 citations

  articleOpen access1st authorCorresponding

  Abstract Many AI development organizations advertise that they have offices of ethics that facilitate ethical AI. However, concerns have been raised that these offices are merely symbolic and do not actually promote ethics. We address the question of how we can know whether an organization is engaging in ethics washing in this way. We articulate an account of organizational power, and we argue that ethics offices that have power are not merely symbolic. Furthermore, we develop a framework for assessing whether an organization has an empowered ethics office—and, thus, is not ethics washing via a symbolic ethics office.

  PublisherOA PDFDOI

• Organizations and Values in Science and Technology

  Philosophy of Science · 2023-10-20 · 3 citations

  articleOpen access1st authorCorresponding

  Abstract This paper articulates a conceptual framework for examining philosophical issues such as the role of values in science at an organizational level. It distinguishes between three dimensions of organizations – organizational aims, organizational structure, and organizational culture – and it examines how these dimensions relate to values in research and development, with a focus on machine learning systems for predictive policing. This framework can be fruitful in identifying interesting and understudied philosophical problems – including those involving inter-organizational divisions of labor – that might otherwise be difficult to conceptualize.

  PublisherOA PDFDOI

• Applications and Societal Implications of Artificial Intelligence in Manufacturing: A Systematic Review

  arXiv (Cornell University) · 2023-07-25 · 6 citations

  reviewOpen access

  This paper undertakes a systematic review of relevant extant literature to consider the potential societal implications of the growth of AI in manufacturing. We analyze the extensive range of AI applications in this domain, such as interfirm logistics coordination, firm procurement management, predictive maintenance, and shop-floor monitoring and control of processes, machinery, and workers. Additionally, we explore the uncertain societal implications of industrial AI, including its impact on the workforce, job upskilling and deskilling, cybersecurity vulnerability, and environmental consequences. After building a typology of AI applications in manufacturing, we highlight the diverse possibilities for AI's implementation at different scales and application types. We discuss the importance of considering AI's implications both for individual firms and for society at large, encompassing economic prosperity, equity, environmental health, and community safety and security. The study finds that there is a predominantly optimistic outlook in prior literature regarding AI's impact on firms, but that there is substantial debate and contention about adverse effects and the nature of AI's societal implications. The paper draws analogies to historical cases and other examples to provide a contextual perspective on potential societal effects of industrial AI. Ultimately, beneficial integration of AI in manufacturing will depend on the choices and priorities of various stakeholders, including firms and their managers and owners, technology developers, civil society organizations, and governments. A broad and balanced awareness of opportunities and risks among stakeholders is vital not only for successful and safe technical implementation but also to construct a socially beneficial and sustainable future for manufacturing in the age of AI.

  PublisherOA PDFDOI

• Global AI Ethics Documents: What They Reveal About Motivations, Practices, and Policies

  ˜The œInternational library of ethics, law and technology · 2022-01-01 · 17 citations

  book-chapter
  PublisherDOI

• Socializing Science: On the Epistemic Significance of the Institutional Context of Science

  Figshare · 2022-09-15 · 6 citations

  article1st authorCorresponding

  Science is a social activity. About this much, everyone agrees. Disagreement arises, however, with the attempt to explicate the precise manner in which science is social and the precise ways in which scientific communities should be organized. This dissertation addresses these two issues. <p> In the first part of this dissertation, I argue that scientific knowledge is intrinsically social; that is, the development of scientific knowledge requires communities, and the decisions that communities make--regarding both the choices of problems and the epistemic evaluation of research--are inevitably influenced by their broader social contexts. In some areas of science, furthermore, moral and political values inevitably influence the epistemic evaluation of research. I defend this claim against the objection that only the practical--and not the epistemic--evaluation of research is legitimately influenced by such values. </p><p> Given that scientific knowledge is intrinsically social, the question of how research should be organized takes on added epistemic significance. If the development of scientific knowledge requires communities, one must begin to spell out which communities are knowledge-productive and which are not. The second part of the dissertation is an attempt to do this, particularly in the area of pharmaceutical research. Due to the increasing involvement of for-profit companies in biomedical research, the organization of this research is changing dramatically, in ways that are presenting serious ethical and epistemic costs. The question of how these organizational arrangements should be improved is not only an important public policy issue, but also an important epistemological one. To begin to address this issue, I discuss a recent episode of pharmaceutical research, involving Vioxx, and I highlight some of the organizational inadequacies that led to this debacle. </p><p> A crucial lesson of the Vioxx case is the need for enforcing organized skepticism. I discuss one potential way of doing this--through an adversarial system of research--according to which competing sets of scientific advocacy groups argue for opposing positions before a panel of scientist-judges. While this proposal is not completely developed, it is a pursuit-worthy one that merits further discussion.</p>

  DOI

Frequent coauthors

• Daniel Schiff

  Purdue University System

  12 shared

• Jason Borenstein

  Georgia Institute of Technology

  12 shared

• Kelly Laas

  12 shared

• Laura North

  2 shared

• Philip Shapira

  2 shared

• Ronan A Lyons

  Swansea University

  2 shared

• John P. Nelson

  2 shared

• Anna Leuschner

  University of Wuppertal

  2 shared

Education

• Ph.D., History and Philosophy of Science

  University of Notre Dame

• M.A., History and Philosophy of Science

  University of Notre Dame

Awards & honors

• Senior Fellow of SOCRATES (“Social Credibility and Trustwort…
• Distinguished Fellow at the Notre Dame Institute for Advance…