Papers in Progress

“HOW SHOULD AI TALK ABOUT US?” Here I explore the ethics of AI speech and generative AI’s depiction of social groups by assessing the ways ChatGPT uses generics. We use generics to track, represent, and communicate information about kinds of entities in the world and the features they seem to share. Social generic statements have the structure of “Ks F,” where K stands for a social category such as “women,” and F stands for purported shared features such as “are nurturing.” Recent research has suggested that simply using generics to summarize regularities can foster essentialist beliefs that group members’ aptitudes and life outcomes are fixed, thus exacerbating barriers to achieving justice. As we increasingly turn to generative artificial intelligence, particularly language models, in our everyday epistemic practices, it is troubling that such technologies are powered by generalizations and communicate via generic statements. After analyzing and categorizing troubling usages, I assess five ethically, psychologically, and technologically informed strategies for improving language models’ use of generics. Instead of avoidance, I advocate for a constructive solution involving narrative contextualization and highlight opportunities to harness the power of generative AI for social change.

“GROUP HYPOCRISY” (under review) In practically every facet of life, we are radically dependent on institutions and, as a result, vulnerable to their failures. Here I diagnose an underexplored source of such failure at the collective level—hypocrisy. In my view, group hypocrisy is a deceptive practice wherein a group avows commitment to one value (or set of values) while its actions are primarily geared toward another value (or set of values). It differs from and is irreducible to pervasive individual hypocrisy occurring in group contexts. Next, I differentiate hypocrisy and two relevant collective phenomena recently theorized by philosophers: untrustworthiness (Fricker, 2023) and deception (Lackey, 2020/2021). I suggest that group hypocrisy is a distinct source of institutional failure and public distrust. I conclude by discussing the urgency for collective accountability and structural interventions to combat group hypocrisy effectively.

“THREE TARGETS OF RELIGIOUS DISCRIMINATION AND THE COMMITMENT ACCOUNT OF COLLECTIVE RELIGIOUS BELIEF” (under review) How do perpetrators of religious discrimination identify their targets? Theorists have distinguished between two targeted groups: (1) religious believers and (2) non-believers who are assumed or perceived to be believers. Members of the second group may be targeted based on attributive features such as ethnicity, skin color, attire, and place of origin. Further, it is theorized that the first group is presumably targeted because of their religious belief. However, who are the subjects of religious belief? Philosophical accounts have primarily assumed that religious believers are just individuals with a doxastic attitude toward propositions relevant to a given religion. Yet it seems implausible that all who are presumed to be believers hold the attitude of belief toward doctrines taken to be central to their religion. Thus, it would be wrong to assume that individuals are primarily or only targeted because of their religious beliefs. It would also be misguided to assume that non-believers are primarily or only targeted based on attributive features. Here, I posit a collectivist, commitment-based account that supplements our understanding of religious belief and believers. On my view, religious believers constitute a conceptually distinct though often overlapping group from religious adherents. Adherents are those members of religious communities who commit to upholding certain religious doctrines and values. They need not, however, have the attitude of belief toward those doctrines or values. Lastly, I raise and respond to the argument from hypocrisy against the commitment account. I contend that my account is compatible with the existence of religious hypocrisy.

“DIAGNOSING EPISTEMIC PATHOLOGIES: COLLECTIVE DISTRUST” (co-authored with Anna Pederneschi) Non-ideal epistemology has been in the business of diagnosing epistemic pathologies, which affect inquirers in their ability to acquire information. In this paper, we diagnose a new epistemic pathology: collective distrust. It occurs when a group of knowers, by their commitment to a community, adopts a distrustful attitude toward new information or sources of information. Collective distrust has the potential to foster an overly vigilant or even paranoid outlook on new information and impede our knowledge practices. We start by defining distrust as an attitude arising from a judgment of untrustworthiness. Distrust is epistemically legitimate when based on evidence and reasons and illegitimate when based on motivated reasoning, namely, when biases interfere with belief acquisition or updating. Moreover, bias-based distrust can spread across domains of interaction and disseminate to entire groups as the object of distrust. At this point, we claim that distrust can be a collective attitude when a group of inquirers commit to adopting a distrustful attitude toward some information or source of information. When this occurs, group members default to the group’s stance rather than their own when evaluating a new piece of information. Next, we argue that collective distrust is an epistemic pathology in two ways. First, it jeopardizes the flow of information and the acquisition of knowledge at the interpersonal level. An individual with a pro-in-group bias irrationally distrusts information from out-group sources, thereby missing pieces of knowledge. This claim is corroborated by the counterfactual that had the inquirer not been part of the group, she would have rationally acquired justification for a new belief. Second, distrust is a pathology at the group level when it becomes the default stance of a group for encountering new information or counterevidence. A default stance of distrust impedes epistemic and practical collaboration and increases group members’ susceptibility to motivated reasoning and other cognitive biases. Finally, we highlight that collective distrust can be epistemically and practically beneficial when it plays a defensive role for systematically and institutionally marginalized groups, whose distrustful stance may be justified in light of their social positioning.

M.A. Thesis

“EXPERIENCE-BASED INTUITIONS” I argue that many identification intuitions, such as one that helps you identify the authorship of a painting you are seeing for the first time, fall under the class of experience-based intuitions. Such identification intuitions cannot arise without intuition generating systems (IGSs) that are shaped by experiences accumulated during one’s life. On my view, experience-based intuitions are produced by domain-general learning systems of hierarchical abstraction which may be modeled by deep convolutional neural networks. Owing to the mechanism of such IGSs, the reliability of experience-based intuition X depends on the quality of the experiences underlying the IGS which produces X. Lastly, I suggest that insofar as some philosophical thought experiments elicit experience-based identification intuitions, we can use the case method to glean information about our experiences as well as uncover certain conceptual commitments.

Affiliations

2023-2025 Graduate Project Fellow, Genericity, Stability, and Structural Interactions Project, UC Irvine and California State University East Bay, Co-PIs Katherine Ritchie and Ny Vasil

This NSF-funded research project at the intersection of philosophy, psychology and cognitive science studies the ways children and adults generalize about the social and natural world. Responsibilities include drafting papers, mentoring undergraduate research assistants, assisting with experimental design, and organizing workshops.

2022-2024 Research Assistant to Margaret Gilbert, UC Irvine

Responsibilities include helping with the development of the forthcoming Palgrave Handbook on the Philosophy of Rights and preparing the indices for Life in Groups (2023).

2022-2024 Graduate Student Associate, Social Unity Research Cluster, UC Irvine

SURC is an interdisciplinary research group that investigates the social and psychological forces that make for social unity and their potential to overcome divisions. Responsibilities include generating relevant research and attending lectures and workshops.

2017 Graduate Laboratory Assistant, Imaging Genetics and Informatics Laboratory, Georgia State University

Responsibilities included designing a study survey on auditory verbal hallucinations and conducting quantitative and qualitative data analysis