Beyond Categories of Caste: Examining Caste Bias and Morality in Text-to-Image AI Models

• URL: http://arxiv.org/abs/2606.00039v1
• Date: Tue, 28 Apr 2026 03:13:11 GMT
• Title: Beyond Categories of Caste: Examining Caste Bias and Morality in Text-to-Image AI Models
• Abstract summary: Recent work has shown caste biases and stereotypes are being perpetuated through Generative AI (GenAI) systems.<n>We shift our ontology to focus on the relational aspect of caste.<n>We propose an anti-caste approach to tackle the issue of caste bias and fairness in AI systems.
• Score: 17.19140224115358
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Text-to-Image (T2I) models have shown promising utility across various domains. However, such models are also amplifying harmful societal biases in their outputs. In the context of South Asia, recent work has shown caste biases and stereotypes are being perpetuated through Generative AI (GenAI) systems. While this research offers extremely relevant insight into invisibilized narratives of caste discrimination through the GenAI system, they often treat caste as an identity category. Therefore, in this work we shift our ontology to focus on the relational aspect of caste. This enables us to develop a more nuanced understanding of the mechanics of caste discrimination by and through T2I models. Combining an algorithmic audit with critical discourse analysis, we draw on a conceptual frame challenging Brahminical Normativity to show how caste biases are perpetuated beyond the simple binaries of upper vs lower-caste categories. Our contributions are two-fold. Beyond challenging the categorical understanding of caste as a category, we propose an anti-caste approach to tackle the issue of caste bias and fairness in AI systems.

Related papers

• Sima AIunty: Caste Audit in LLM-Driven Matchmaking [10.466211034402386]
  We conduct a controlled audit of caste bias in large language models (LLMs)-mediated matchmaking evaluations.<n>We vary caste identity across Brahmin, Kshatriya, Vaishya, Shudra, and Dalit, and income across five buckets.<n>Our analysis reveals consistent hierarchical patterns across models: same-caste matches are rated most favorably, with average ratings up to 25% higher than inter-caste matches.
  arXiv  Detail & Related papers  (2026-03-31T05:44:55Z)
• DECASTE: Unveiling Caste Stereotypes in Large Language Models through Multi-Dimensional Bias Analysis [20.36241144630387]
  Large language models (LLMs) have revolutionized natural language processing (NLP)<n>LLMs have been shown to reflect and perpetuate harmful societal biases, including those based on ethnicity, gender, and religion.<n>We propose DECASTE, a novel framework designed to detect and assess both implicit and explicit caste biases in LLMs.
  arXiv  Detail & Related papers  (2025-05-20T23:19:13Z)
• SelEx: Self-Expertise in Fine-Grained Generalized Category Discovery [55.72840638180451]
  Generalized Category Discovery aims to simultaneously uncover novel categories and accurately classify known ones. Traditional methods, which lean heavily on self-supervision and contrastive learning, often fall short when distinguishing between fine-grained categories. We introduce a novel concept called self-expertise', which enhances the model's ability to recognize subtle differences and uncover unknown categories.
  arXiv  Detail & Related papers  (2024-08-26T15:53:50Z)
• Interpretations, Representations, and Stereotypes of Caste within Text-to-Image Generators [0.0]
  This paper addresses interpretations, representations, and stereotypes surrounding a tragically underexplored context in T2I research: caste. We examine how the T2I Stable Diffusion displays people of various castes, and what professions they are depicted as performing. Our findings reveal how Stable Diffusion outputs perpetuate systems of 'castelessness', equating Indianness with high-castes and depicting caste-oppressed identities with markers of poverty.
  arXiv  Detail & Related papers  (2024-08-02T22:06:41Z)
• Generative Discrimination: What Happens When Generative AI Exhibits Bias, and What Can Be Done About It [2.2913283036871865]
  chapter explores how genAI intersects with non-discrimination laws. It highlights two main types of discriminatory outputs: (i) demeaning and abusive content and (ii) subtler biases due to inadequate representation of protected groups. It argues for holding genAI providers and deployers liable for discriminatory outputs and highlights the inadequacy of traditional legal frameworks to address genAI-specific issues.
  arXiv  Detail & Related papers  (2024-06-26T13:32:58Z)
• Hacking a surrogate model approach to XAI [49.1574468325115]
  We show that even if a discriminated subgroup does not get a positive decision from the black box ADM system, the corresponding question of group membership can be pushed down onto a level as low as wanted. Our approach can be generalized easily to other surrogate models.
  arXiv  Detail & Related papers  (2024-06-24T13:18:02Z)
• The Devil is in the Neurons: Interpreting and Mitigating Social Biases in Pre-trained Language Models [78.69526166193236]
  Pre-trained Language models (PLMs) have been acknowledged to contain harmful information, such as social biases. We propose sc Social Bias Neurons to accurately pinpoint units (i.e., neurons) in a language model that can be attributed to undesirable behavior, such as social bias. As measured by prior metrics from StereoSet, our model achieves a higher degree of fairness while maintaining language modeling ability with low cost.
  arXiv  Detail & Related papers  (2024-06-14T15:41:06Z)
• The Male CEO and the Female Assistant: Evaluation and Mitigation of Gender Biases in Text-To-Image Generation of Dual Subjects [58.27353205269664]
  We propose the Paired Stereotype Test (PST) framework, which queries T2I models to depict two individuals assigned with male-stereotyped and female-stereotyped social identities.<n>PST queries T2I models to depict two individuals assigned with male-stereotyped and female-stereotyped social identities.<n>Using PST, we evaluate two aspects of gender biases -- the well-known bias in gendered occupation and a novel aspect: bias in organizational power.
  arXiv  Detail & Related papers  (2024-02-16T21:32:27Z)
• Stable Bias: Analyzing Societal Representations in Diffusion Models [72.27121528451528]
  We propose a new method for exploring the social biases in Text-to-Image (TTI) systems. Our approach relies on characterizing the variation in generated images triggered by enumerating gender and ethnicity markers in the prompts. We leverage this method to analyze images generated by 3 popular TTI systems and find that while all of their outputs show correlations with US labor demographics, they also consistently under-represent marginalized identities to different extents.
  arXiv  Detail & Related papers  (2023-03-20T19:32:49Z)
• Factoring the Matrix of Domination: A Critical Review and Reimagination of Intersectionality in AI Fairness [55.037030060643126]
  Intersectionality is a critical framework that allows us to examine how social inequalities persist. We argue that adopting intersectionality as an analytical framework is pivotal to effectively operationalizing fairness.
  arXiv  Detail & Related papers  (2023-03-16T21:02:09Z)
• Explaining Caste-based Digital Divide in India [0.0]
  This paper analyzes the first-level digital divide (ownership of computer and access to the internet) and the second-level digital divide (individual's skill to use computer and the internet) between the disadvantaged caste group and the others. Results indicate that the caste-based digital divide in India is rooted in historical socioeconomic deprivation of disadvantaged caste groups.
  arXiv  Detail & Related papers  (2021-06-30T09:19:24Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.