William S. Agnew

Organizers Of QueerInAI, Anaelia Ovalle, Arjun Subramonian et al. · allen-ai, cmu

We present Queer in AI as a case study for community-led participatory design in AI. We examine how participatory design and intersectional tenets started and shaped this community's programs over the years. We discuss different challenges that emerged in the process, look at ways this organization has fallen short of operationalizing participatory and intersectional principles, and then assess the organization's impact. Queer in AI provides important lessons and insights for practitioners and theorists of participatory methods broadly through its rejection of hierarchy in favor of decentralization, success at building aid and programs by and for the queer community, and effort to change actors and institutions outside of the queer community. Finally, we theorize how communities like Queer in AI contribute to the participatory design in AI more broadly by fostering cultures of participation in AI, welcoming and empowering marginalized participants, critiquing poor or exploitative participatory practices, and bringing participation to institutions outside of individual research projects. Queer in AI's work serves as a case study of grassroots activism and participatory methods within AI, demonstrating the potential of community-led participatory methods and intersectional praxis, while also providing challenges, case studies, and nuanced insights to researchers developing and using participatory methods.

Irene Solaiman, Zeerak Talat, William Agnew et al. · allen-ai, cmu

Generative AI systems across modalities, ranging from text (including code), image, audio, and video, have broad social impacts, but there is no official standard for means of evaluating those impacts or for which impacts should be evaluated. In this paper, we present a guide that moves toward a standard approach in evaluating a base generative AI system for any modality in two overarching categories: what can be evaluated in a base system independent of context and what can be evaluated in a societal context. Importantly, this refers to base systems that have no predetermined application or deployment context, including a model itself, as well as system components, such as training data. Our framework for a base system defines seven categories of social impact: bias, stereotypes, and representational harms; cultural values and sensitive content; disparate performance; privacy and data protection; financial costs; environmental costs; and data and content moderation labor costs. Suggested methods for evaluation apply to listed generative modalities and analyses of the limitations of existing evaluations serve as a starting point for necessary investment in future evaluations. We offer five overarching categories for what can be evaluated in a broader societal context, each with its own subcategories: trustworthiness and autonomy; inequality, marginalization, and violence; concentration of authority; labor and creativity; and ecosystem and environment. Each subcategory includes recommendations for mitigating harm.

Organizers of QueerInAI, Nathan Dennler, Anaelia Ovalle et al. · allen-ai, meta-ai

Bias evaluation benchmarks and dataset and model documentation have emerged as central processes for assessing the biases and harms of artificial intelligence (AI) systems. However, these auditing processes have been criticized for their failure to integrate the knowledge of marginalized communities and consider the power dynamics between auditors and the communities. Consequently, modes of bias evaluation have been proposed that engage impacted communities in identifying and assessing the harms of AI systems (e.g., bias bounties). Even so, asking what marginalized communities want from such auditing processes has been neglected. In this paper, we ask queer communities for their positions on, and desires from, auditing processes. To this end, we organized a participatory workshop to critique and redesign bias bounties from queer perspectives. We found that when given space, the scope of feedback from workshop participants goes far beyond what bias bounties afford, with participants questioning the ownership, incentives, and efficacy of bounties. We conclude by advocating for community ownership of bounties and complementing bounties with participatory processes (e.g., co-creation).

Andrew Hundt, William Agnew, Vicky Zeng et al. · cmu

Stereotypes, bias, and discrimination have been extensively documented in Machine Learning (ML) methods such as Computer Vision (CV) [18, 80], Natural Language Processing (NLP) [6], or both, in the case of large image and caption models such as OpenAI CLIP [14]. In this paper, we evaluate how ML bias manifests in robots that physically and autonomously act within the world. We audit one of several recently published CLIP-powered robotic manipulation methods, presenting it with objects that have pictures of human faces on the surface which vary across race and gender, alongside task descriptions that contain terms associated with common stereotypes. Our experiments definitively show robots acting out toxic stereotypes with respect to gender, race, and scientifically-discredited physiognomy, at scale. Furthermore, the audited methods are less likely to recognize Women and People of Color. Our interdisciplinary sociotechnical analysis synthesizes across fields and applications such as Science Technology and Society (STS), Critical Studies, History, Safety, Robotics, and AI. We find that robots powered by large datasets and Dissolution Models (sometimes called "foundation models", e.g. CLIP) that contain humans risk physically amplifying malignant stereotypes in general; and that merely correcting disparities will be insufficient for the complexity and scale of the problem. Instead, we recommend that robot learning methods that physically manifest stereotypes or other harmful outcomes be paused, reworked, or even wound down when appropriate, until outcomes can be proven safe, effective, and just. Finally, we discuss comprehensive policy changes and the potential of new interdisciplinary research on topics like Identity Safety Assessment Frameworks and Design Justice to better understand and address these harms.

Sabine Weber, Angelina Wang, Ankush Gupta et al. · meta-ai

Natural language processing (NLP) technologies are rapidly reshaping how language is created, processed, and interpreted by humans. With current and potential applications in hiring, law, healthcare, and other areas that impact people's lives, understanding and mitigating harms towards marginalized groups is critical. In this survey, we examine NLP research papers that explicitly address the relationship between LGBTQIA+ communities and NLP technologies. We systematically review all such papers published in the ACL Anthology up until February 2026 (n=122), to answer the following research questions: (1) What are current research trends? (2) What gaps exist in terms of topics and methods? (3) What areas are open for future work? We find that while the number of papers on queer NLP has grown within the last few years, most papers take a reactive rather than a proactive approach, focusing on shortcomings of existing systems rather than creating new solutions. Our survey uncovers many opportunities for future work, especially regarding stakeholder involvement, intersectionality, interdisciplinarity, and languages other than English. We also offer an outlook from a queer studies perspective, highlighting understudied topics and blind spots in the harms addressed in NLP papers. Beyond being a roadmap of what has been done, this survey is a call to action for work towards more just and inclusive NLP technologies.

Pratyusha Ria Kalluri, William Agnew, Myra Cheng et al. · allen-ai

A rapidly growing number of voices argue that AI research, and computer vision in particular, is powering mass surveillance. Yet the direct path from computer vision research to surveillance has remained obscured and difficult to assess. Here, we reveal the Surveillance AI pipeline by analyzing three decades of computer vision research papers and downstream patents, more than 40,000 documents. We find the large majority of annotated computer vision papers and patents self-report their technology enables extracting data about humans. Moreover, the majority of these technologies specifically enable extracting data about human bodies and body parts. We present both quantitative and rich qualitative analysis illuminating these practices of human data extraction. Studying the roots of this pipeline, we find that institutions that prolifically produce computer vision research, namely elite universities and "big tech" corporations, are subsequently cited in thousands of surveillance patents. Further, we find consistent evidence against the narrative that only these few rogue entities are contributing to surveillance. Rather, we expose the fieldwide norm that when an institution, nation, or subfield authors computer vision papers with downstream patents, the majority of these papers are used in surveillance patents. In total, we find the number of papers with downstream surveillance patents increased more than five-fold between the 1990s and the 2010s, with computer vision research now having been used in more than 11,000 surveillance patents. Finally, in addition to the high levels of surveillance we find documented in computer vision papers and patents, we unearth pervasive patterns of documents using language that obfuscates the extent of surveillance. Our analysis reveals the pipeline by which computer vision research has powered the ongoing expansion of surveillance.

Rachel Hong, Yael Eiger, Jevan Hutson et al.

Pluralistic alignment has emerged as a promising approach for ensuring that large language models (LLMs) faithfully represent the diversity, nuance, and conflict inherent in human values. In this work, we study a high-stakes deployment context - mulching - where automated systems transform selected individuals into nutrient-rich slurry for the dual purposes of food security and aesthetic population management. Building on recent pluralistic alignment frameworks, we introduce ValueMulch, a reproducible training, deployment, and certification pipeline for aligning mulching models (MMs) to a wide range of community norms. Through a real-world testbed spanning 32 communities, we show that ValueMulch improves distributional agreement with community mulching preferences relative to frontier baselines. We conclude with a discussion of ethical considerations, limitations, and implications for researchers seeking to align systems to the full spectrum of human values - especially when those values are inconsistent, commercially inconvenient, or nutritionally underutilized. Author's note: This piece builds on prior existing work Keyes et al in 2019 that satirized cannibalism as a parody for approaches that imbue ethics into problematic technology. We bring those ideas to today's era with the proliferation of large language models in everyday lives, as a critique of current AI pluralistic alignment literature. Our work does not intend to argue that all alignment practices are evil, but rather that if framing value design as a technical problem enables technology systems to enact harms, then perhaps this framing is not enough.

William Agnew, Harry H. Jiang, Cella Sum et al. · allen-ai, cmu

Large language models excel at performing inference over text to extract information, summarize information, or generate additional text. These inference capabilities are implicated in a variety of ethical harms spanning surveillance, labor displacement, and IP/copyright theft. While many policy, legal, and technical mitigations have been proposed to counteract these harms, these mitigations typically require cooperation from institutions that move slower than technical advances (i.e., governments) or that have few incentives to act to counteract these harms (i.e., the corporations that create and profit from these LLMs). In this paper, we define and build "data defenses" -- a novel strategy that directly empowers data owners to block LLMs from performing inference on their data. We create data defenses by developing a method to automatically generate adversarial prompt injections that, when added to input text, significantly reduce the ability of LLMs to accurately infer personally identifying information about the subject of the input text or to use copyrighted text in inference. We examine the ethics of enabling such direct resistance to LLM inference, and argue that making data defenses that resist and subvert LLMs enables the realization of important values such as data ownership, data sovereignty, and democratic control over AI systems. We verify that our data defenses are cheap and fast to generate, work on the latest commercial and open-source LLMs, resistance to countermeasures, and are robust to several different attack settings. Finally, we consider the security implications of LLM data defenses and outline several future research directions in this area. Our code is available at https://github.com/wagnew3/LLMDataDefenses and a tool for using our defenses to protect text against LLM inference is at https://wagnew3.github.io/LLM-Data-Defenses/.

William Agnew, Christopher Xie, Aaron Walsman et al.

Learning-based 3D object reconstruction enables single- or few-shot estimation of 3D object models. For robotics, this holds the potential to allow model-based methods to rapidly adapt to novel objects and scenes. Existing 3D reconstruction techniques optimize for visual reconstruction fidelity, typically measured by chamfer distance or voxel IOU. We find that when applied to realistic, cluttered robotics environments, these systems produce reconstructions with low physical realism, resulting in poor task performance when used for model-based control. We propose ARM, an amodal 3D reconstruction system that introduces (1) a stability prior over object shapes, (2) a connectivity prior, and (3) a multi-channel input representation that allows for reasoning over relationships between groups of objects. By using these priors over the physical properties of objects, our system improves reconstruction quality not just by standard visual metrics, but also performance of model-based control on a variety of robotics manipulation tasks in challenging, cluttered environments. Code is available at github.com/wagnew3/ARM.

Jared Moore, Declan Grabb, William Agnew et al.

Should a large language model (LLM) be used as a therapist? In this paper, we investigate the use of LLMs to *replace* mental health providers, a use case promoted in the tech startup and research space. We conduct a mapping review of therapy guides used by major medical institutions to identify crucial aspects of therapeutic relationships, such as the importance of a therapeutic alliance between therapist and client. We then assess the ability of LLMs to reproduce and adhere to these aspects of therapeutic relationships by conducting several experiments investigating the responses of current LLMs, such as `gpt-4o`. Contrary to best practices in the medical community, LLMs 1) express stigma toward those with mental health conditions and 2) respond inappropriately to certain common (and critical) conditions in naturalistic therapy settings -- e.g., LLMs encourage clients' delusional thinking, likely due to their sycophancy. This occurs even with larger and newer LLMs, indicating that current safety practices may not address these gaps. Furthermore, we note foundational and practical barriers to the adoption of LLMs as therapists, such as that a therapeutic alliance requires human characteristics (e.g., identity and stakes). For these reasons, we conclude that LLMs should not replace therapists, and we discuss alternative roles for LLMs in clinical therapy.

Chung Peng Lee, Rachel Hong, Harry Jiang et al.

The internet has become the main source of data to train modern text-to-image or vision-language models, yet it is increasingly unclear whether web-scale data collection practices for training AI systems adequately respect data owners' wishes. Ignoring the owner's indication of consent around data usage not only raises ethical concerns but also has recently been elevated into lawsuits around copyright infringement cases. In this work, we aim to reveal information about data owners' consent to AI scraping and training, and study how it's expressed in DataComp, a popular dataset of 12.8 billion text-image pairs. We examine both the sample-level information, including the copyright notice, watermarking, and metadata, and the web-domain-level information, such as a site's Terms of Service (ToS) and Robots Exclusion Protocol. We estimate at least 122M of samples exhibit some indication of copyright notice in CommonPool, and find that 60\% of the samples in the top 50 domains come from websites with ToS that prohibit scraping. Furthermore, we estimate 9-13\% with 95\% confidence interval of samples from CommonPool to contain watermarks, where existing watermark detection methods fail to capture them in high fidelity. Our holistic methods and findings show that data owners rely on various channels to convey data consent, of which current AI data collection pipelines do not entirely respect. These findings highlight the limitations of the current dataset curation/release practice and the need for a unified data consent framework taking AI purposes into consideration.

William Agnew, Julia Barnett, Annie Chu et al.

Generative audio models are rapidly advancing in both capabilities and public utilization -- several powerful generative audio models have readily available open weights, and some tech companies have released high quality generative audio products. Yet, while prior work has enumerated many ethical issues stemming from the data on which generative visual and textual models have been trained, we have little understanding of similar issues with generative audio datasets, including those related to bias, toxicity, and intellectual property. To bridge this gap, we conducted a literature review of hundreds of audio datasets and selected seven of the most prominent to audit in more detail. We found that these datasets are biased against women, contain toxic stereotypes about marginalized communities, and contain significant amounts of copyrighted work. To enable artists to see if they are in popular audio datasets and facilitate exploration of the contents of these datasets, we developed a web tool audio datasets exploration tool at https://audio-audit.vercel.app.

Martin Mundt, Anaelia Ovalle, Felix Friedrich et al.

In a widely popular analogy by Turing Award Laureate Yann LeCun, machine intelligence has been compared to cake - where unsupervised learning forms the base, supervised learning adds the icing, and reinforcement learning is the cherry on top. We expand this 'cake that is intelligence' analogy from a simple structural metaphor to the full life-cycle of AI systems, extending it to sourcing of ingredients (data), conception of recipes (instructions), the baking process (training), and the tasting and selling of the cake (evaluation and distribution). Leveraging our re-conceptualization, we describe each step's entailed social ramifications and how they are bounded by statistical assumptions within machine learning. Whereas these technical foundations and social impacts are deeply intertwined, they are often studied in isolation, creating barriers that restrict meaningful participation. Our re-conceptualization paves the way to bridge this gap by mapping where technical foundations interact with social outcomes, highlighting opportunities for cross-disciplinary dialogue. Finally, we conclude with actionable recommendations at each stage of the metaphorical AI cake's life-cycle, empowering prospective AI practitioners, users, and researchers, with increased awareness and ability to engage in broader AI discourse.

Jordan Taylor, William Agnew, Maarten Sap et al.

Visual generative AI models are trained using a one-size-fits-all measure of aesthetic appeal. However, what is deemed "aesthetic" is inextricably linked to personal taste and cultural values, raising the question of whose taste is represented in visual generative AI models. In this work, we study an aesthetic evaluation model--LAION Aesthetic Predictor (LAP)--that is widely used to curate datasets to train visual generative image models, like Stable Diffusion, and evaluate the quality of AI-generated images. To understand what LAP measures, we audited the model across three datasets. First, we examined the impact of aesthetic filtering on the LAION-Aesthetics Dataset (approximately 1.2B images), which was curated from LAION-5B using LAP. We find that the LAP disproportionally filters in images with captions mentioning women, while filtering out images with captions mentioning men or LGBTQ+ people. Then, we used LAP to score approximately 330k images across two art datasets, finding the model rates realistic images of landscapes, cityscapes, and portraits from western and Japanese artists most highly. In doing so, the algorithmic gaze of this aesthetic evaluation model reinforces the imperial and male gazes found within western art history. In order to understand where these biases may have originated, we performed a digital ethnography of public materials related to the creation of LAP. We find that the development of LAP reflects the biases we found in our audits, such as the aesthetic scores used to train LAP primarily coming from English-speaking photographers and western AI-enthusiasts. In response, we discuss how aesthetic evaluation can perpetuate representational harms and call on AI developers to shift away from prescriptive measures of "aesthetics" toward more pluralistic evaluation.

Ashwin, William Agnew, Umut Pajaro et al.

Trustworthy artificial intelligence (AI) has become an important topic because trust in AI systems and their creators has been lost. Researchers, corporations, and governments have long and painful histories of excluding marginalized groups from technology development, deployment, and oversight. As a result, these technologies are less useful and even harmful to minoritized groups. We argue that any AI development, deployment, and monitoring framework that aspires to trust must incorporate both feminist, non-exploitative participatory design principles and strong, outside, and continual monitoring and testing. We additionally explain the importance of considering aspects of trustworthiness beyond just transparency, fairness, and accountability, specifically, to consider justice and shifting power to the disempowered as core values to any trustworthy AI system. Creating trustworthy AI starts by funding, supporting, and empowering grassroots organizations like Queer in AI so the field of AI has the diversity and inclusion to credibly and effectively develop trustworthy AI. We leverage the expert knowledge Queer in AI has developed through its years of work and advocacy to discuss if and how gender, sexuality, and other aspects of queer identity should be used in datasets and AI systems and how harms along these lines should be mitigated. Based on this, we share a gendered approach to AI and further propose a queer epistemology and analyze the benefits it can bring to AI. We additionally discuss how to regulate AI with this queer epistemology in vision, proposing frameworks for making policies related to AI & gender diversity and privacy & queer data protection.

Abeba Birhane, Pratyusha Kalluri, Dallas Card et al.

Machine learning currently exerts an outsized influence on the world, increasingly affecting institutional practices and impacted communities. It is therefore critical that we question vague conceptions of the field as value-neutral or universally beneficial, and investigate what specific values the field is advancing. In this paper, we first introduce a method and annotation scheme for studying the values encoded in documents such as research papers. Applying the scheme, we analyze 100 highly cited machine learning papers published at premier machine learning conferences, ICML and NeurIPS. We annotate key features of papers which reveal their values: their justification for their choice of project, which attributes of their project they uplift, their consideration of potential negative consequences, and their institutional affiliations and funding sources. We find that few of the papers justify how their project connects to a societal need (15\%) and far fewer discuss negative potential (1\%). Through line-by-line content analysis, we identify 59 values that are uplifted in ML research, and, of these, we find that the papers most frequently justify and assess themselves based on Performance, Generalization, Quantitative evidence, Efficiency, Building on past work, and Novelty. We present extensive textual evidence and identify key themes in the definitions and operationalization of these values. Notably, we find systematic textual evidence that these top values are being defined and applied with assumptions and implications generally supporting the centralization of power.Finally, we find increasingly close ties between these highly cited papers and tech companies and elite universities.

Jesse Dodge, Maarten Sap, Ana Marasović et al.

Large language models have led to remarkable progress on many NLP tasks, and researchers are turning to ever-larger text corpora to train them. Some of the largest corpora available are made by scraping significant portions of the internet, and are frequently introduced with only minimal documentation. In this work we provide some of the first documentation for the Colossal Clean Crawled Corpus (C4; Raffel et al., 2020), a dataset created by applying a set of filters to a single snapshot of Common Crawl. We begin by investigating where the data came from, and find a significant amount of text from unexpected sources like patents and US military websites. Then we explore the content of the text itself, and find machine-generated text (e.g., from machine translation systems) and evaluation examples from other benchmark NLP datasets. To understand the impact of the filters applied to create this dataset, we evaluate the text that was removed, and show that blocklist filtering disproportionately removes text from and about minority individuals. Finally, we conclude with some recommendations for how to created and document web-scale datasets from a scrape of the internet.

William Agnew, Pedro Domingos

Current deep reinforcement learning (RL) approaches incorporate minimal prior knowledge about the environment, limiting computational and sample efficiency. \textit{Objects} provide a succinct and causal description of the world, and many recent works have proposed unsupervised object representation learning using priors and losses over static object properties like visual consistency. However, object dynamics and interactions are also critical cues for objectness. In this paper we propose a framework for reasoning about object dynamics and behavior to rapidly determine minimal and task-specific object representations. To demonstrate the need to reason over object behavior and dynamics, we introduce a suite of RGBD MuJoCo object collection and avoidance tasks that, while intuitive and visually simple, confound state-of-the-art unsupervised object representation learning algorithms. We also highlight the potential of this framework on several Atari games, using our object representation and standard RL and planning algorithms to learn dramatically faster than existing deep RL algorithms.