Unlike with traditional predictive AI models, today’s generative AI models are increasingly designed to be general-purpose, able to perform a wide range of tasks. This makes it challenging to develop a reliable and useful understanding of the ways in which this technology is and could be used. As a result, academic and policy researchers and generative AI providers have started to publish the results of their own investigations about the use of generative AI. This information is, however, fragmented, potentially incomplete, sometimes ambiguous, and often lacking in methodological specificity. In this paper, we conducted an integrative review to build a multi-dimensional framework that specifies what kind of information about generative AI use could be reported and how, and illustrated its analytical utility by applying the framework to a collection of over 110 industry documents. Our analysis reveals systematic patterns and omissions in current industry reporting and reflects on the narratives this reporting collectively advance about generative AI use.

Large language models (LLMs) have been positioned to revolutionize HCI, by reshaping not only the interfaces, design patterns, and sociotechnical systems that we study, but also the research practices we use.To-date, however, there has been little understanding of LLMs' uptake in HCI.We address this gap via a systematic literature review of 153 CHI papers from 2020-24 that engage with LLMs.We taxonomize: (1) domains where LLMs are applied; (2) roles of LLMs in HCI projects; (3) contribution types; and (4) acknowledged limitations and risks.We find LLM work in 10 diverse domains, primarily via empirical and artifact contributions.Authors use LLMs in five distinct roles, including as research tools or simulated users.Still, authors often raise validity and reproducibility concerns, and overwhelmingly study closed models.We outline opportunities to improve HCI research with and on LLMs, and provide guiding questions for researchers to consider the validity and appropriateness of LLM-related work.

In this position paper we argue that conclusions drawn about relative system safety or attack method efficacy via AI red teaming are often not supported by evidence provided by attack success rate (ASR) comparisons. We show, through conceptual, theoretical, and empirical contributions, that many conclusions are founded on apples-to-oranges comparisons or low-validity measurements. Our arguments are grounded in asking a simple question: When can attack success rates be meaningfully compared? To answer this question, we draw on ideas from social science measurement theory and inferential statistics, which, taken together, provide a conceptual grounding for understanding when numerical values obtained through the quantification of system attributes can be meaningfully compared. Through this lens, we articulate conditions under which ASRs can and cannot be meaningfully compared. Using jailbreaking as a running example, we provide examples and extensive discussion of apples-to-oranges ASR comparisons and measurement validity challenges.

"Machine unlearning" is a popular proposed solution for mitigating the existence of content in an AI model that is problematic for legal or moral reasons, including privacy, copyright, safety, and more. For example, unlearning is often invoked as a solution for removing the effects of specific information from a generative-AI model's parameters, e.g., a particular individual's personal data or the inclusion of copyrighted content in the model's training data. Unlearning is also proposed as a way to prevent a model from generating targeted types of information in its outputs, e.g., generations that closely resemble a particular individual's data or reflect the concept of "Spiderman." Both of these goals--the targeted removal of information from a model and the targeted suppression of information from a model's outputs--present various technical and substantive challenges. We provide a framework for ML researchers and policymakers to think rigorously about these challenges, identifying several mismatches between the goals of unlearning and feasible implementations. These mismatches explain why unlearning is not a general-purpose solution for circumscribing generative-AI model behavior in service of broader positive impact.

The LLM-as-a-judge paradigm, in which a judge LLM system replaces human raters in rating the outputs of other generative AI (GenAI) systems, plays a critical role in scaling and standardizing GenAI evaluations. To validate such judge systems, evaluators assess human--judge agreement by first collecting multiple human ratings for each item in a validation corpus, then aggregating the ratings into a single, per-item gold label rating. For many items, however, rating criteria may admit multiple valid interpretations, so a human or LLM rater may deem multiple ratings "reasonable" or "correct." We call this condition rating indeterminacy. Problematically, many rating tasks that contain rating indeterminacy rely on forced-choice elicitation, whereby raters are instructed to select only one rating for each item. In this paper, we introduce a framework for validating LLM-as-a-judge systems under rating indeterminacy. We draw theoretical connections between different measures of judge system performance under different human--judge agreement metrics, and different rating elicitation and aggregation schemes. We demonstrate that differences in how humans and LLMs resolve rating indeterminacy when responding to forced-choice rating instructions can heavily bias LLM-as-a-judge validation. Through extensive experiments involving 11 real-world rating tasks and 9 commercial LLMs, we show that standard validation approaches that rely upon forced-choice ratings select judge systems that are highly suboptimal, performing as much as 31% worse than judge systems selected by our approach that uses multi-label "response set" ratings to account for rating indeterminacy. We conclude with concrete recommendations for more principled approaches to LLM-as-a-judge validation.

In AI research and practice, rigor remains largely understood in terms of methodological rigor---such as whether mathematical, statistical, or computational methods are correctly applied. We argue that this narrow conception of rigor has contributed to the concerns raised by the responsible AI community, including overblown claims about the capabilities of AI systems. Our position is that a broader conception of what rigorous AI research and practice should entail is needed. We believe such a conception---in addition to a more expansive understanding of 1) methodological rigor---should include aspects related to 2) what background knowledge informs what to work on (epistemic rigor); 3) how disciplinary, community, or personal norms, standards, or beliefs influence the work (normative rigor); 4) how clearly articulated the theoretical constructs under use are (conceptual rigor); 5) what is reported and how (reporting rigor); and 6) how well-supported the inferences from existing evidence are (interpretative rigor). In doing so, we also provide useful language and a framework for much needed dialogue about the AI community's work by researchers, policymakers, journalists, and other stakeholders.

As AI technologies improve, people are increasingly willing to delegate tasks to AI agents. In many cases, the human decision-maker chooses whether to delegate to an AI agent based on properties of the specific instance of the decision-making problem they are facing. Since humans typically lack full awareness of all the factors relevant to this choice for a given decision-making instance, they perform a kind of categorization by treating indistinguishable instances - those that have the same observable features - as the same. In this paper, we define the problem of designing the optimal algorithmic delegate in the presence of categories. This is an important dimension in the design of algorithms to work with humans, since we show that the optimal delegate can be an arbitrarily better teammate than the optimal standalone algorithmic agent. The solution to this optimal delegation problem is not obvious: we discover that this problem is fundamentally combinatorial, and illustrate the complex relationship between the optimal design and the properties of the decision-making task even in simple settings. Indeed, we show that finding the optimal delegate is computationally hard in general. However, we are able to find efficient algorithms for producing the optimal delegate in several broad cases of the problem, including when the optimal action may be decomposed into functions of features observed by the human and the algorithm. Finally, we run computational experiments to simulate a designer updating an algorithmic delegate over time to be optimized for when it is actually adopted by users, and show that while this process does not recover the optimal delegate in general, the resulting delegate often performs quite well.

Recent years have witnessed increasing calls for computing researchers to grapple with the societal impacts of their work. Tools such as impact assessments have gained prominence as a method to uncover potential impacts, and a number of publication venues now encourage authors to include an impact statement in their submissions. Despite this recent push, little is known about the way researchers go about assessing, articulating, and addressing the potential negative societal impact of their work --- especially in industry settings, where research outcomes are often quickly integrated into products and services. In addition, while there are nascent efforts to support researchers in this task, there remains a dearth of empirically-informed tools and processes. Through interviews with 25 industry computing researchers across different companies and research areas, we identify four key factors that influence how they grapple with (or choose not to grapple with) the societal impact of their research: the relationship between industry researchers and product teams; organizational dynamics and cultures that prioritize innovation and speed; misconceptions about societal impact; and a lack of sufficient infrastructure to support researchers. To develop an effective impact assessment template tailored to industry computing researchers' needs, we conduct an iterative co-design process with these 25 industry researchers, along with an additional 16 researchers and practitioners with prior experience and expertise in reviewing and developing impact assessments or responsible computing practices more broadly. Through the co-design process, we develop 10 design considerations to facilitate the effective design, implementation, and adaptation of an impact assessment template for use in industry research settings and beyond, as well as our own "Societal Impact Assessment" template with concrete scaffolds. We explore the effectiveness of this template through a user study with 15 industry research interns, revealing both its strengths and limitations. Finally, we discuss the implications for future researchers, organizations, and policymakers seeking to foster more responsible research practices.

Variance in predictions across different trained models is a significant, under-explored source of error in fair binary classification. In practice, the variance on some data examples is so large that decisions can be effectively arbitrary. To investigate this problem, we take an experimental approach and make four overarching contributions. We: 1) Define a metric called self-consistency, derived from variance, which we use as a proxy for measuring and reducing arbitrariness; 2) Develop an ensembling algorithm that abstains from classification when a prediction would be arbitrary; 3) Conduct the largest to-date empirical study of the role of variance (vis-a-vis self-consistency and arbitrariness) in fair binary classification; and, 4) Release a toolkit that makes the US Home Mortgage Disclosure Act (HMDA) datasets easily usable for future research. Altogether, our experiments reveal shocking insights about the reliability of conclusions on benchmark datasets. Most fair binary classification benchmarks are close-to-fair when taking into account the amount of arbitrariness present in predictions -- before we even try to apply any fairness interventions. This finding calls into question the practical utility of common algorithmic fairness methods, and in turn suggests that we should reconsider how we choose to measure fairness in binary classification.

Predicting future outcomes is a prevalent application of machine learning in social impact domains. Examples range from predicting student success in education to predicting disease risk in healthcare. Practitioners recognize that the ultimate goal is not just to predict but to act effectively. Increasing evidence suggests that relying on outcome predictions for downstream interventions may not have desired results. In most domains there exists a multitude of possible interventions for each individual, making the challenge of taking effective action more acute. Even when causal mechanisms connecting the individual's latent states to outcomes are well understood, in any given instance (a specific student or patient), practitioners still need to infer---from budgeted measurements of latent states---which of many possible interventions will be most effective for this individual. With this in mind, we ask: when are accurate predictors of outcomes helpful for identifying the most suitable intervention? Through a simple model encompassing actions, latent states, and measurements, we demonstrate that pure outcome prediction rarely results in the most effective policy for taking actions, even when combined with other measurements. We find that except in cases where there is a single decisive action for improving the outcome, outcome prediction never maximizes "action value", the utility of taking actions. Making measurements of actionable latent states, where specific actions lead to desired outcomes, may considerably enhance the action value compared to outcome prediction, and the degree of improvement depends on action costs and the outcome model. This analysis emphasizes the need to go beyond generic outcome prediction in interventional settings by incorporating knowledge of plausible actions and latent states.

In this paper, we examine computational approaches for measuring the "fairness" of image tagging systems, finding that they cluster into five distinct categories, each with its own analytic foundation. We also identify a range of normative concerns that are often collapsed under the terms "unfairness," "bias," or even "discrimination" when discussing problematic cases of image tagging. Specifically, we identify four types of representational harms that can be caused by image tagging systems, providing concrete examples of each. We then consider how different computational measurement approaches map to each of these types, demonstrating that there is not a one-to-one mapping. Our findings emphasize that no single measurement approach will be definitive and that it is not possible to infer from the use of a particular measurement approach which type of harm was intended to be measured. Lastly, equipped with this more granular understanding of the types of representational harms that can be caused by image tagging systems, we show that attempts to mitigate some of these types of harms may be in tension with one another.

An emerging theme in artificial intelligence research is the creation of models to simulate the decisions and behavior of specific people, in domains including game-playing, text generation, and artistic expression. These models go beyond earlier approaches in the way they are tailored to individuals, and the way they are designed for interaction rather than simply the reproduction of fixed, pre-computed behaviors. We refer to these as mimetic models, and in this paper we develop a framework for characterizing the ethical and social issues raised by their growing availability. Our framework includes a number of distinct scenarios for the use of such models, and considers the impacts on a range of different participants, including the target being modeled, the operator who deploys the model, and the entities that interact with it.

Despite growing calls for participation in AI design, there are to date few empirical studies of what these processes look like and how they can be structured for meaningful engagement with domain experts. In this paper, we examine a notable yet understudied AI design process in the legal domain that took place over a decade ago, the impact of which still informs legal automation efforts today. Specifically, we examine the design and evaluation activities that took place from 2006 to 2011 within the Text REtrieval Conference's (TREC) Legal Track, a computational research venue hosted by the National Institute of Standards and Technologies. The Legal Track of TREC is notable in the history of AI research and practice because it relied on a range of participatory approaches to facilitate the design and evaluation of new computational techniques-in this case, for automating attorney document review for civil litigation matters. Drawing on archival research and interviews with coordinators of the Legal Track of TREC, our analysis reveals how an interactive simulation methodology allowed computer scientists and lawyers to become co-designers and helped bridge the chasm between computational research and real-world, high-stakes litigation practice. In analyzing this case from the recent past, our aim is to empirically ground contemporary critiques of AI development and evaluation and the calls for greater participation as a means to address them.

Disaggregated evaluations of AI systems, in which system performance is assessed and reported separately for different groups of people, are conceptually simple. However, their design involves a variety of choices. Some of these choices influence the results that will be obtained, and thus the conclusions that can be drawn; others influence the impacts---both beneficial and harmful---that a disaggregated evaluation will have on people, including the people whose data is used to conduct the evaluation. We argue that a deeper understanding of these choices will enable researchers and practitioners to design careful and conclusive disaggregated evaluations. We also argue that better documentation of these choices, along with the underlying considerations and tradeoffs that have been made, will help others when interpreting an evaluation's results and conclusions.

Scholars have recently drawn attention to a range of controversial issues posed by the use of computer vision for automatically generating descriptions of people in images. Despite these concerns, automated image description has become an important tool to ensure equitable access to information for blind and low vision people. In this paper, we investigate the ethical dilemmas faced by companies that have adopted the use of computer vision for producing alt text: textual descriptions of images for blind and low vision people. We use Facebook's automatic alt text tool as our primary case study. First, we analyze the policies that Facebook has adopted with respect to identity categories, such as race, gender, age, etc., and the company's decisions about whether to present these terms in alt text. We then describe an alternative---and manual---approach practiced in the museum community, focusing on how museums determine what to include in alt text descriptions of cultural artifacts. We compare these policies, using notable points of contrast to develop an analytic framework that characterizes the particular apprehensions behind these policy choices. We conclude by considering two strategies that seem to sidestep some of these concerns, finding that there are no easy ways to avoid the normative dilemmas posed by the use of computer vision to automate alt text.

Many policies allocate harms or benefits that are uncertain in nature: they produce distributions over the population in which individuals have different probabilities of incurring harm or benefit. Comparing different policies thus involves a comparison of their corresponding probability distributions, and we observe that in many instances the policies selected in practice are hard to explain by preferences based only on the expected value of the total harm or benefit they produce. In cases where the expected value analysis is not a sufficient explanatory framework, what would be a reasonable model for societal preferences over these distributions? Here we investigate explanations based on the framework of probability weighting from the behavioral sciences, which over several decades has identified systematic biases in how people perceive probabilities. We show that probability weighting can be used to make predictions about preferences over probabilistic distributions of harm and benefit that function quite differently from expected-value analysis, and in a number of cases provide potential explanations for policy preferences that appear hard to motivate by other means. In particular, we identify optimal policies for minimizing perceived total harm and maximizing perceived total benefit that take the distorting effects of probability weighting into account, and we discuss a number of real-world policies that resemble such allocational strategies. Our analysis does not provide specific recommendations for policy choices, but is instead fundamentally interpretive in nature, seeking to describe observed phenomena in policy choices.

We survey 146 papers analyzing "bias" in NLP systems, fnding that their motivations are often vague, inconsistent, and lacking in normative reasoning, despite the fact that analyzing "bias" is an inherently normative process.We further fnd that these papers' proposed quantitative techniques for measuring or mitigating "bias" are poorly matched to their motivations and do not engage with the relevant literature outside of NLP.Based on these fndings, we describe the beginnings of a path forward by proposing three recommendations that should guide work analyzing "bias" in NLP systems.These recommendations rest on a greater recognition of the relationships between language and social hierarchies, encouraging researchers and practitioners to articulate their conceptualizations of "bias"-i.e., what kinds of system behaviors are harmful, in what ways, to whom, and why, as well as the normative reasoning underlying these statements-and to center work around the lived experiences of members of communities affected by NLP systems, while interrogating and reimagining the power relations between technologists and such communities.NLP task Papers Embeddings (type-level or contextualized) 54 Coreference resolution 20 Language modeling or dialogue generation 17 Hate-speech detection 17 Sentiment analysis 15 Machine translation 8 Tagging or parsing 5 Surveys, frameworks, and meta-analyses 20 Other 22

Designing technical systems to be resistant to bias and discrimination represents vital new terrain for researchers, policymakers, and the anti-discrimination project more broadly. We consider bias and discrimination in the context of popular online dating and hookup platforms in the United States, which we call intimate platforms. Drawing on work in social-justice-oriented and Queer HCI, we review design features of popular intimate platforms and their potential role in exacerbating or mitigating interpersonal bias. We argue that focusing on platform design can reveal opportunities to reshape troubling patterns of intimate contact without overriding users' decisional autonomy. We identify and address the difficult ethical questions that nevertheless come along with such intervention, while urging the social computing community to engage more deeply with issues of bias, discrimination, and exclusion in the study and design of intimate platforms.

The use of big data research methods has grown tremendously over the past five years in both academia and industry. As the size and complexity of available datasets has grown, so too have the ethical questions raised by big data research. These questions become increasingly urgent as data and research agendas move well beyond those typical of the computational and natural sciences, to more directly address sensitive aspects of human behavior, interaction, and health. The tools of big data research are increasingly woven into our daily lives, including mining digital medical records for scientific and economic insights, mapping relationships via social media, capturing individuals’ speech and action via sensors, tracking movement across space, shaping police and security policy via “predictive policing,” and much more.

No abstract available.