Emily Denton
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Emily Denton has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
Remi Denton
Google Research, USA
5 shared publications- Eric Corbett
Google Research, New York, New York, USA
2 shared publications Alex Hanna
Google Research, NY, USA
1 shared publication- Alex S. Taylor
University of Edinburgh
1 shared publication Andrew Smart
Google Research, NY, USA
1 shared publication- Casey Fiesler
University of Colorado, Boulder, Colorado, USA
1 shared publication
Latest Publications
- Open Access
- 30 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Power and Public Participation in AI
The rapid growth of AI in contemporary life has outpaced the public participation necessary for society to determine how these technologies should be used. As scholars respond to this challenge by exploring new modes of public participation in AI, we help advance these efforts by introducing influential work from public planning scholarship, the Ladder of Citizen Participation, as an analytical lens to help compare and contrast power in this work. We used the ladder to analyze participatory approaches to AI development in recent scholarship, finding that most of this work informs or consults rather than partners with or delegates control to participants. We also found that papers frequently reflect a writing style that makes it difficult to ascertain the degree of power afforded. We discuss implications from our work for powerholders (developers, researchers, practitioners) offering participatory approaches to AI and for people (specific communities, stakeholders, general public) participating in those processes.
- 30 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Open AccessPower and Public Participation in AI
The rapid growth of AI in contemporary life has outpaced the public participation necessary for society to determine how these technologies should be used. As scholars respond to this challenge by exploring new modes of public participation in AI, we help advance these efforts by introducing influential work from public planning scholarship, the Ladder of Citizen Participation, as an analytical lens to help compare and contrast power in this work. We used the ladder to analyze participatory approaches to AI development in recent scholarship, finding that most of this work informs or consults rather than partners with or delegates control to participants. We also found that papers frequently reflect a writing style that makes it difficult to ascertain the degree of power afforded. We discuss implications from our work for powerholders (developers, researchers, practitioners) offering participatory approaches to AI and for people (specific communities, stakeholders, general public) participating in those processes.
- Top Cited PapersOpen Access
- 12 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 506-517
AI’s Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia
This paper presents a community-centered study of cultural limitations of text-to-image (T2I) models in the South Asian context. We theorize these failures using scholarship on dominant media regimes of representations and locate them within participants’ reporting of their existing social marginalizations. We thus show how generative AI can reproduce an outsiders gaze for viewing South Asian cultures, shaped by global and regional power inequities. By centering communities as experts and soliciting their perspectives on T2I limitations, our study adds rich nuance into existing evaluative frameworks and deepens our understanding of the culturally-specific ways AI technologies can fail in non-Western and Global South settings. We distill lessons for responsible development of T2I models, recommending concrete pathways forward that can allow for recognition of structural inequalities.
- 12 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 506-517
Top Cited PapersOpen AccessAI’s Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia
This paper presents a community-centered study of cultural limitations of text-to-image (T2I) models in the South Asian context. We theorize these failures using scholarship on dominant media regimes of representations and locate them within participants’ reporting of their existing social marginalizations. We thus show how generative AI can reproduce an outsiders gaze for viewing South Asian cultures, shaped by global and regional power inequities. By centering communities as experts and soliciting their perspectives on T2I limitations, our study adds rich nuance into existing evaluative frameworks and deepens our understanding of the culturally-specific ways AI technologies can fail in non-Western and Global South settings. We distill lessons for responsible development of T2I models, recommending concrete pathways forward that can allow for recognition of structural inequalities.
- Open Access
- 12 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1624-1634
Interrogating the T in FAccT
Fairness, accountability, and transparency are the three conceptual foundations of the FAccT conference. Transparency, however, has yet to be scrutinized to the same degree as accountability and fairness. As a result, we don't know: What does this community mean when it talks about transparency? How are we doing transparency? And to what ends? What commitments does (or should) the T in FAccT signify? This paper interrogates the T in FAccT using perspectives from critical transparency literature. Subsequently, we argue that FAccT might be better off dropping the T from its title for two reasons: (1) transparency can often be counterproductive to FAccT's primary objectives and (2) it is misleading as FAccT is mainly preoccupied with explainability rather than actual transparency. If we want to keep the T, we need to reframe how we think about and do transparency by making transparency contingent, reclaiming it from explainability, and bringing people into transparency processes.
- 12 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1624-1634
Open AccessInterrogating the T in FAccT
Fairness, accountability, and transparency are the three conceptual foundations of the FAccT conference. Transparency, however, has yet to be scrutinized to the same degree as accountability and fairness. As a result, we don't know: What does this community mean when it talks about transparency? How are we doing transparency? And to what ends? What commitments does (or should) the T in FAccT signify? This paper interrogates the T in FAccT using perspectives from critical transparency literature. Subsequently, we argue that FAccT might be better off dropping the T from its title for two reasons: (1) transparency can often be counterproductive to FAccT's primary objectives and (2) it is misleading as FAccT is mainly preoccupied with explainability rather than actual transparency. If we want to keep the T, we need to reframe how we think about and do transparency by making transparency contingent, reclaiming it from explainability, and bringing people into transparency processes.
- Top Cited PapersOpen Access
- 12 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 205-216
"I wouldn't say offensive but...": Disability-Centered Perspectives on Large Language Models
Large language models (LLMs) trained on real-world data can inadvertently reflect harmful societal biases, particularly toward historically marginalized communities. While previous work has primarily focused on harms related to age and race, emerging research has shown that biases toward disabled communities exist. This study extends prior work exploring the existence of harms by identifying categories of LLM-perpetuated harms toward the disability community. We conducted 19 focus groups, during which 56 participants with disabilities probed a dialog model about disability and discussed and annotated its responses. Participants rarely characterized model outputs as blatantly offensive or toxic. Instead, participants used nuanced language to detail how the dialog model mirrored subtle yet harmful stereotypes they encountered in their lives and dominant media, e.g., inspiration porn and able-bodied saviors. Participants often implicated training data as a cause for these stereotypes and recommended training the model on diverse identities from disability-positive resources. Our discussion further explores representative data strategies to mitigate harm related to different communities through annotation co-design with ML researchers and developers.
- 12 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 205-216
Top Cited PapersOpen Access"I wouldn't say offensive but...": Disability-Centered Perspectives on Large Language Models
Large language models (LLMs) trained on real-world data can inadvertently reflect harmful societal biases, particularly toward historically marginalized communities. While previous work has primarily focused on harms related to age and race, emerging research has shown that biases toward disabled communities exist. This study extends prior work exploring the existence of harms by identifying categories of LLM-perpetuated harms toward the disability community. We conducted 19 focus groups, during which 56 participants with disabilities probed a dialog model about disability and discussed and annotated its responses. Participants rarely characterized model outputs as blatantly offensive or toxic. Instead, participants used nuanced language to detail how the dialog model mirrored subtle yet harmful stereotypes they encountered in their lives and dominant media, e.g., inspiration porn and able-bodied saviors. Participants often implicated training data as a cause for these stereotypes and recommended training the model on diverse identities from disability-positive resources. Our discussion further explores representative data strategies to mitigate harm related to different communities through annotation co-design with ML researchers and developers.
- Open Access
- 14 April 2023
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 7 (CSCW1) , 1-33
From Human to Data to Dataset: Mapping the Traceability of Human Subjects in Computer Vision Datasets
Computer vision is a "data hungry" field. Researchers and practitioners who work on human-centric computer vision, like facial recognition, emphasize the necessity of vast amounts of data for more robust and accurate models. Humans are seen as a data resource which can be converted into datasets. The necessity of data has led to a proliferation of gathering data from easily available sources, including "public" data from the web. Yet the use of public data has significant ethical implications for the human subjects in datasets. We bridge academic conversations on the ethics of using publicly obtained data with concerns about privacy and agency associated with computer vision applications. Specifically, we examine how practices of dataset construction from public data-not only from websites, but also from public settings and public records-make it extremely difficult for human subjects to trace their images as they are collected, converted into datasets, distributed for use, and, in some cases, retracted. We discuss two interconnected barriers current data practices present to providing an ethics of traceability for human subjects: awareness and control. We conclude with key intervention points for enabling traceability for data subjects. We also offer suggestions for an improved ethics of traceability to enable both awareness and control for individual subjects in dataset curation practices.
- 14 April 2023
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 7 (CSCW1) , 1-33
Open AccessFrom Human to Data to Dataset: Mapping the Traceability of Human Subjects in Computer Vision Datasets
Computer vision is a "data hungry" field. Researchers and practitioners who work on human-centric computer vision, like facial recognition, emphasize the necessity of vast amounts of data for more robust and accurate models. Humans are seen as a data resource which can be converted into datasets. The necessity of data has led to a proliferation of gathering data from easily available sources, including "public" data from the web. Yet the use of public data has significant ethical implications for the human subjects in datasets. We bridge academic conversations on the ethics of using publicly obtained data with concerns about privacy and agency associated with computer vision applications. Specifically, we examine how practices of dataset construction from public data-not only from websites, but also from public settings and public records-make it extremely difficult for human subjects to trace their images as they are collected, converted into datasets, distributed for use, and, in some cases, retracted. We discuss two interconnected barriers current data practices present to providing an ethics of traceability for human subjects: awareness and control. We conclude with key intervention points for enabling traceability for data subjects. We also offer suggestions for an improved ethics of traceability to enable both awareness and control for individual subjects in dataset curation practices.
- Top Cited PapersOpen Access
- 1 July 2021
- Published by SAGE Publications in Big Data & Society
- Vol. 8 (2)
On the genealogy of machine learning datasets: A critical history of ImageNet
In response to growing concerns of bias, discrimination, and unfairness perpetuated by algorithmic systems, the datasets used to train and evaluate machine learning models have come under increased scrutiny. Many of these examinations have focused on the contents of machine learning datasets, finding glaring underrepresentation of minoritized groups. In contrast, relatively little work has been done to examine the norms, values, and assumptions embedded in these datasets. In this work, we conceptualize machine learning datasets as a type of informational infrastructure, and motivate a genealogy as method in examining the histories and modes of constitution at play in their creation. We present a critical history of ImageNet as an exemplar, utilizing critical discourse analysis of major texts around ImageNet’s creation and impact. We find that assumptions around ImageNet and other large computer vision datasets more generally rely on three themes: the aggregation and accumulation of more data, the computational construction of meaning, and making certain types of data labor invisible. By tracing the discourses that surround this influential benchmark, we contribute to the ongoing development of the standards and norms around data development in machine learning and artificial intelligence research.
- 1 July 2021
- Published by SAGE Publications in Big Data & Society
- Vol. 8 (2)
Top Cited PapersOpen AccessOn the genealogy of machine learning datasets: A critical history of ImageNet
In response to growing concerns of bias, discrimination, and unfairness perpetuated by algorithmic systems, the datasets used to train and evaluate machine learning models have come under increased scrutiny. Many of these examinations have focused on the contents of machine learning datasets, finding glaring underrepresentation of minoritized groups. In contrast, relatively little work has been done to examine the norms, values, and assumptions embedded in these datasets. In this work, we conceptualize machine learning datasets as a type of informational infrastructure, and motivate a genealogy as method in examining the histories and modes of constitution at play in their creation. We present a critical history of ImageNet as an exemplar, utilizing critical discourse analysis of major texts around ImageNet’s creation and impact. We find that assumptions around ImageNet and other large computer vision datasets more generally rely on three themes: the aggregation and accumulation of more data, the computational construction of meaning, and making certain types of data labor invisible. By tracing the discourses that surround this influential benchmark, we contribute to the ongoing development of the standards and norms around data development in machine learning and artificial intelligence research.