Towards Fairness in Practice: A Practitioner-Oriented Rubric for Evaluating Fair ML Toolkits

Richardson, Brianna; Garcia-Gathright, Jean; Way, Samuel F.; Thom, Jennifer S.; Cramer, Henriette

doi:10.1145/3411764.3445604

Cited by 34 publications

(25 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We searched in two ways. First, we looked at references from recent research papers from FAccT and CHI that survey ethical toolkits [e.g., 33,54]. Second, following the approach in Lee and Singh [33], we emulated the position of a practitioner looking for ethical toolkits and conducted a range of Google searches for artifacts using the terms: "AI ethics toolkit, " "AI values toolkit, " "AI fairness toolkit, " "ethics design toolkit, " "values design toolkit. "…”

Section: Corpus Developmentmentioning

confidence: 99%

“…2.1.2 In AI ethics. In light of AI practitioners' needs for support in addressing the ethical dimensions of AI [22], technology companies, researchers at FAccT, AIES, CHI, and other venues, and others have developed numerous tools and resources to support that work, with many such resources taking the form of toolkits [17,31,33,41,43,54,60]. Several papers have performed systemic meta-reviews and empirical analyses of AI ethics toolkits [5,33,43,54].…”

Section: Introductionmentioning

confidence: 99%

“…In light of AI practitioners' needs for support in addressing the ethical dimensions of AI [22], technology companies, researchers at FAccT, AIES, CHI, and other venues, and others have developed numerous tools and resources to support that work, with many such resources taking the form of toolkits [17,31,33,41,43,54,60]. Several papers have performed systemic meta-reviews and empirical analyses of AI ethics toolkits [5,33,43,54]. Ayling and Chapman [5] perform a descriptive analysis of AI ethics toolkits, identifying stakeholder types common across toolkits, and stages in the organizational lifecycle at which various toolkits are applied.…”

Section: Introductionmentioning

confidence: 99%

“…Lee and Singh [33] take a normative look at six open source fairness toolkits, using surveys and interviews with practitioners to understand the strengths and weaknesses of these tools. Richardson et al [54] performed a simulated ethics scenario with ML practitioners, observing their experience using various ethics toolkits to inform recommendations for their design. Morley et al [43] proposes a typology of AI ethics approaches condensed from a variety of toolkits, also critiquing toolkits' lack of usability and their focus on individuals rather than social groups.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Seeing Like a Toolkit: How Toolkits Envision the Work of AI Ethics

Wong¹,

Madaio²,

Merrill³

2022

Preprint

View full text Add to dashboard Cite

Numerous toolkits have been developed to support ethical AI development. However, toolkits, like all tools, encode assumptions in their design about what work should be done and how. In this paper, we conduct a qualitative analysis of 27 AI ethics toolkits to critically examine how the work of ethics is imagined and how it is supported by these toolkits. Specifically, we examine the discourses toolkits rely on when talking about ethical issues, who they imagine should do the work of ethics, and how they envision the work practices involved in addressing ethics. We find that AI ethics toolkits largely frame the work of AI ethics to be technical work for individual technical practitioners, despite calls for engaging broader sets of stakeholders in grappling with social aspects of AI ethics, and without contending with the organizational and political implications of AI ethics work in practice. Among all toolkits, we identify a mismatch between the imagined work of ethics and the support the toolkits provide for doing that work. We identify a lack of guidance around how to navigate organizational power dynamics as they relate to performing ethical work. We use these omissions to chart future work for researchers and designers of AI ethics toolkits.CCS Concepts: • Social and professional topics → Codes of ethics; Computing occupations; Computing organizations.

show abstract

Section: Corpus Developmentmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Seeing Like a Toolkit: How Toolkits Envision the Work of AI Ethics

Wong¹,

Madaio²,

Merrill³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In order to explore practitioners' perceptions and desires around open-source fairness toolkits, Lee et al conducted interview studies and a survey to identify the gaps between the capabilities of existing fairness toolkits and the needs of industry practitioners [63]. In a similar vein, Richardson et al conducted an interview study with twenty ML practitioners in a simulated scenario in order to generate a practitioner-oriented rubric for evaluating fair ML toolkits [89]. However, neither of these two works engaged practitioners directly in using a fairness toolkit within the context of a real ML task.…”

Section: Introductionmentioning

confidence: 99%

Exploring How Machine Learning Practitioners (Try To) Use Fairness Toolkits

Deng,

Nagireddy,

Lee

et al. 2022

Preprint

View full text Add to dashboard Cite

Recent years have seen the development of many open-source ML fairness toolkits aimed at helping ML practitioners assess and address unfairness in their systems. However, there has been little research investigating how ML practitioners actually use these toolkits in practice. In this paper, we conducted the first in-depth empirical exploration of how industry practitioners (try to) work with existing fairness toolkits. In particular, we conducted think-aloud interviews to understand how participants learn about and use fairness toolkits, and explored the generality of our findings through an anonymous online survey. We identified several opportunities for fairness toolkits to better address practitioner needs and scaffold them in using toolkits effectively and responsibly. Based on these findings, we highlight implications for the design of future open-source fairness toolkits that can support practitioners in better contextualizing, communicating, and collaborating around ML fairness efforts.

show abstract

The AI field needs translational Ethical AI research

Borg

2022

AI Magazine

View full text Add to dashboard Cite

Calls for Ethical AI have become urgent and pervasive, especially as ethical issues surrounding AI products at tech companies are increasingly scrutinized by the public. Yet even after a first wave of responses to these calls coalesced around Ethical AI principles to guide decision-making and a second wave generated technical tools to mitigate specific ethical issues, multiple lines of evidence indicate that these Ethical AI principles and technical tools have only a limited impact on the daily practices of AI users and producers. In other words, there is a big gap between what we publish in academic papers and what AI creators need to generate AI products that reflect society's values. Ethical AI is by no means the only field to have this problem. However, when medical and ecology fields documented similar gaps between their fields' scientific discoveries and the practices and products that people actually use, they invested tremendous resources into subfields that developed evidence about how to translate what was done in the lab to adopted solutions. I argue in this commentary that it is our research community's moral duty to invest in our own subfield of "Translational Ethical AI" that will determine how best to ensure AI practitioners can implement the Ethical AI technical tools we publish in academic venues in production settings. Further, I offer concrete steps for doing that, drawing on insights gleaned from other translational fields. Closing the "Ethical AI Publication-to-Practice gap" will be a considerable transdisciplinary challenge, but one of the AI research community has the unique expertise, political leverage, and moral responsibility to tackle.

show abstract

Towards Fairness in Practice: A Practitioner-Oriented Rubric for Evaluating Fair ML Toolkits

Cited by 34 publications

References 23 publications

Seeing Like a Toolkit: How Toolkits Envision the Work of AI Ethics

Seeing Like a Toolkit: How Toolkits Envision the Work of AI Ethics

Exploring How Machine Learning Practitioners (Try To) Use Fairness Toolkits

The AI field needs translational Ethical AI research

Contact Info

Product

Resources

About