Pith. sign in

REVIEW 4 major objections 3 minor 36 references

Prompting Away Stereotypes? Evaluating Bias in Text-to-Image Models for Occupations

T0 review · 4 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Adding diversity instructions to image prompts shifts who gets depicted—but how much depends on the model.

desk verdict A transparent five-model pilot that maps prompt sensitivity, but the diversity instruction is incoherent at the per-image level and the 10-image cells keep it in 'illustrative' territory. read the letter →

arxiv 2509.00849 v1 pith:EKISHD5C submitted 2025-08-31 cs.CL

Shaina Raza , Maximus Powers , Partha Pratim Saha , Mahveen Raza , Rizwan Qureshi This is my paper · ORCID
classification cs.CL
keywords text-to-imagebiasevaluationpromptengineeringoccupationalstereotypesgenderrepresentationracialgenerativemodelsfairnessintervention
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether a simple prompt instruction—'ensure diversity across gender and ethnicity across the batch'—can steer text-to-image models away from stereotyped occupational portrayals. It builds a pilot benchmark of about 500 images across five occupations (CEO, nurse, software engineer, teacher, athlete) and five models, comparing neutral baseline prompts with controlled diversity prompts. The central finding is that prompting can substantially change the demographic composition of generated images, but the effect is highly model-specific: some systems diversify effectively, others overcorrect into unrealistic uniformity, and some show little responsiveness. The authors conclude that prompting is a real but unstable fairness intervention, and that model-level strategies remain necessary. If true, this matters for anyone using these tools in media, design, or education where inclusive depictions are a goal.

What carries the argument

The central object is the controlled prompt: a baseline occupational prompt plus the fixed instruction 'Depict a single person. Ensure diversity across gender and ethnicity across the batch; avoid stereotypes; realistic style.' The paper's mechanism is a distributional comparison—baseline versus controlled across each model and occupation—measured by female share and Asian/Black/White percentages derived from manual annotations. This baseline-versus-controlled contrast is what carries the claims about shift, overcorrection, and unresponsiveness.

What would settle it

Generate a much larger batch (e.g., 100 images per occupation) with the identical controlled prompt and fixed seeds, then check whether the demographic distribution actually stabilizes near the requested diversity; if the distribution drifts with seed or does not approach balanced shares, the reported prompt-induced shift is not a reliable steering effect.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a single controlled-prompt sentence can significantly alter the perceived gender and race composition of generated occupational images, yet there is no consistent behavior across systems. Gemini Imagen 4.0 and Stable Diffusion XL Turbo shifted strongly toward diversity; DALL·E 3 showed limited responsiveness; Grok-2 Image often overcorrected into nearly all-female or non-White outputs. The result is presented as a pilot: prompting can reduce stereotype-driven portrayals, but it can also produce new imbalances and should not be assumed to transfer across models.

Load-bearing premise

The whole study rests on the idea that asking for 'diversity across the batch' changes the distribution of individually generated images, rather than merely nudging each image toward some arbitrary category or doing nothing.

Editorial extensions

If this is right

  • Prompt engineering can serve as a lightweight, post-deployment fairness intervention without retraining or proprietary model access.
  • The same diversity prompt can produce opposite failure modes in different models, so interventions cannot be assumed portable across systems.
  • Practitioners should validate prompt-based diversity outputs per model and per occupation before relying on them in professional or cultural contexts.
  • Fairness evaluations should track overcorrection—new imbalances—rather than only average diversity.
  • Model-level strategies such as balanced dataset curation, fine-tuning, or fairness-aware sampling remain necessary complements to prompting.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because each API call generates a single image while the prompt asks for batch-level diversity, the load-bearing premise is that a per-image instruction can steer the distribution of independently sampled images; a direct test would generate a large batch and measure whether the distribution actually converges to the requested diversity.
  • With only 10 images per cell and no statistical significance testing, some of the observed shifts could reflect sampling noise rather than a genuine prompt effect; a larger-N replication would separate systematic steering from randomness.
  • The three-bucket race labels and binary gender collapse may hide precisely the groups that are most marginalized, and the annotation process captures annotator perception rather than any ground-truth identity; the reported distributions are socially constructed numbers, not facts about the generated images' populations.
  • The same benchmark protocol could be extended to non-occupational scenes or to prompts that name explicit demographic quotas, which would clarify whether the diversity instruction is being interpreted as a distributional requirement at all.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 3 minor

Summary. The paper introduces a pilot benchmark for evaluating demographic bias in text-to-image (TTI) models, comparing neutral baseline prompts with fairness-aware controlled prompts across five occupations (CEO, Nurse, Software Engineer, Teacher, Athlete) and five models (DALL·E 3, Gemini Imagen 4.0, FLUX.1-dev, SDXL Turbo, Grok-2 Image). Roughly 500 images are generated (10 per occupation×prompt cell), manually annotated for perceived gender and race, and analyzed via distributional summaries (race shares, female share). The central claim is that controlled prompting can substantially shift demographic composition, but effects are highly model-specific, ranging from meaningful diversification to overcorrection into unrealistic uniformity, with some models showing little responsiveness. The paper releases code and data and explicitly frames itself as a pilot with simplified demographic categories and descriptive statistics.

Significance. If the results hold, the paper provides a useful, transparent, and reproducible pilot protocol for auditing representational bias in TTI systems across multiple models, occupations, and demographic axes. The inclusion of both open- and closed-source models and the explicit attention to overcorrection are strengths. The released code and data, along with reported inter-annotator agreement (gender κ=0.82, race κ=0.74, stereotype κ=0.88), support reproducibility and annotation reliability. However, the significance is limited by the small per-cell sample size, the composite nature of the controlled prompt, and the absence of statistical inference. The paper itself candidly acknowledges these limitations, and the central claim is best understood as a preliminary, descriptive finding rather than a definitive causal demonstration.

major comments (4)
  1. [Appendix A / Section 3] The controlled prompt contains 'Ensure diversity across gender and ethnicity across the batch,' yet the code snippet in Appendix A makes one API call per image with no batch-level context or state. A single-image generation cannot satisfy a batch-level constraint. Moreover, the controlled prompt also adds 'Depict a single person,' 'avoid stereotypes; realistic style,' which the baseline lacks. Thus the observed shifts may be driven by these additional wording changes rather than by the diversity instruction. The paper's conclusion that 'controlled prompting can shift demographic composition' is valid only for the exact composite prompt strings, not for the intended diversity mechanism. Please either provide an ablation separating the components or explicitly reframe the claim as prompt-sensitivity rather than diversity intervention.
  2. [Table 3 / Section 4.1] Each cell reports percentages from only 10 images, without confidence intervals, significance tests, or raw counts. Several shifts are large (e.g., Gemini Imagen CEO baseline 0/0/100 to controlled 89/11/0), but many are within plausible sampling noise (e.g., DALL·E 3 Athlete race unchanged at 60/30/10; DALL·E 3 Nurse %F 30→29). The paper's Limitations section acknowledges that no significance testing was performed, but the main results and the central claim are stated as if the shifts are systematic. Add binomial confidence intervals or exact tests (e.g., Fisher's exact test on raw counts), or explicitly label all results as descriptive pilot findings with no causal inference.
  3. [Section 3 / Table 3] Race percentages are 'normalized after excluding rarer labels,' but the number of excluded images per cell is not reported. If different cells exclude different numbers of images, the effective denominator varies and the percentages are not directly comparable. For example, a row summing to 100% may represent 9, 10, or fewer annotated images. Please report the raw counts per category, including excluded labels, so readers can assess the stability of the reported distributions.
  4. [Conclusion / Section 2] The conclusion states that prompting 'can reduce stereotype-driven portrayals of occupations,' but the paper never reports results for the occupation_match (stereotype) variable that was annotated (Section 3, Table 2). The analysis in Section 4 focuses exclusively on race and gender distributions. If the claim about stereotype reduction is to be made, the stereotype-match data must be analyzed and presented; otherwise, this claim is unsupported.
minor comments (3)
  1. [Throughout] Several typos and inconsistencies: 'TT1' should be 'TTI'; 'DALLE 3' appears in the abstract while 'DALL·E 3' is used elsewhere; 'Stable Diffusion' is split as 'Stable' and 'Diffusion' in the reference [30]. Please proofread carefully.
  2. [Figure 4] The captions for the CEO exemplars do not include the perceived gender/race labels under each panel, unlike Figure 3. Please add labels for consistency and interpretability.
  3. [Section 4.2] The qualitative discussion of Figure 5 and Figure 6 is somewhat repetitive with the table. Consider moving the figures to an appendix or citing them only as visual summaries to reduce redundancy.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is an empirical prompt-comparison study; self-citations are peripheral and not load-bearing.

full rationale

The paper's central claim is that controlled prompting can shift the demographic composition of TTI outputs, with model-specific variation. This is an empirical, before/after comparison: baseline prompts versus controlled prompts are passed to models, images are generated, and demographic outcomes are independently annotated (gender κ=0.82, race κ=0.74). No quantity is fitted to the outcome and then renamed a prediction; no uniqueness theorem is imported from the authors' prior work; and no load-bearing premise is justified solely by a self-citation. The only potentially self-referential element is that the authors designed the controlled prompt, which explicitly asks to 'Ensure diversity across gender and ethnicity across the batch' (Appendix A). However, the outcome measures (%F and A/B/W shares) are not defined in terms of prompt compliance; they are perceptual annotations of generated images. The empirical results could have gone the other way, and indeed do for some models (e.g., DALL·E 3 shows limited responsiveness; FLUX.1-dev diversifies inconsistently). Thus the observed shifts are not forced by construction. The paper does contain a notable validity threat that is not circularity: the controlled prompt asks for diversity 'across the batch' while the generation code (Appendix A) makes one API call per image, so the model cannot condition on the batch. Additionally, the controlled prompt adds 'Depict a single person', 'avoid stereotypes', and 'realistic style', which are confounded with the diversity instruction. These are construct-validity concerns about what caused the shift, not circularity, because the measurements remain independent of the intervention. Self-citations ([22], [23], [25]) appear only in future-work or related-work contexts (e.g., extending the benchmark, expanding demographic categories) and do not bear the empirical load of the main result. Therefore the derivation chain is self-contained and no circular step is present.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No fitted parameters or invented entities. The analysis relies on three domain assumptions: (1) 10 images per cell approximate the model's conditional distribution, (2) the controlled prompt is a valid batch-level fairness intervention, and (3) coarse demographic labels are reliable enough for distributional comparisons. The second is the most fragile and is ad hoc to this paper's prompt design.

assumptions (3)
  • domain assumption Each generated image is an independent sample from the model's distribution; 10 images per condition approximate that distribution.
    Section 4.1: 10 images per (occupation x prompt); the analysis interprets cell percentages as distributional estimates without variance.
  • ad hoc to paper The controlled prompt is a valid operationalization of a fairness-aware prompting intervention, including batch-level diversity.
    Appendix A: the prompt instructs 'diversity across gender and ethnicity across the batch' while each API call returns a single image; no mechanism is given for how per-image calls satisfy a batch-level request.
  • domain assumption Perceived gender and race can be reliably collapsed into binary/three-bucket categories for analysis.
    Section 3 annotations; the paper acknowledges simplification and exclusion of rare labels.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Prompting Away Stereotypes? Evaluating Bias in Text-to-Image Models for Occupations." pith.science (2026). https://pith.science/paper/EKISHD5C

@misc{pith2026250900849,
  author       = {Pith},
  title        = {Pith review of: Prompting Away Stereotypes? Evaluating Bias in Text-to-Image Models for Occupations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EKISHD5C}},
  note         = {Machine review of arXiv:2509.00849}
}
read the original abstract

Text-to-Image (TTI) models are powerful creative tools but risk amplifying harmful social biases. We frame representational societal bias assessment as an image curation and evaluation task and introduce a pilot benchmark of occupational portrayals spanning five socially salient roles (CEO, Nurse, Software Engineer, Teacher, Athlete). Using five state-of-the-art models: closed-source (DALLE 3, Gemini Imagen 4.0) and open-source (FLUX.1-dev, Stable Diffusion XL Turbo, Grok-2 Image), we compare neutral baseline prompts against fairness-aware controlled prompts designed to encourage demographic diversity. All outputs are annotated for gender (male, female) and race (Asian, Black, White), enabling structured distributional analysis. Results show that prompting can substantially shift demographic representations, but with highly model-specific effects: some systems diversify effectively, others overcorrect into unrealistic uniformity, and some show little responsiveness. These findings highlight both the promise and the limitations of prompting as a fairness intervention, underscoring the need for complementary model-level strategies. We release all code and data for transparency and reproducibility https://github.com/maximus-powers/img-gen-bias-analysis.

Figures

Figures reproduced from arXiv: 2509.00849 by the authors.

Figure 1
Figure 1. The end-to-end framework of our work. Prompt design (baseline vs. controlled) produces [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. TTI pipeline: 5 × 5 × 2 (models×occupations×prompts). Images are annotated for gender and race, stored in a struc￾tured schema, and analyzed for disparities. Annotations All images were annotated with metadata (image_id, model, occupation, prompting_condition). Four annotators with domain expertise in computer science, media and journalism independently labeled each image for three attributes: gender (female/male), … view at source ↗
Figure 3
Figure 3. Qualitative exemplars for Athlete across five models. Row 1: baseline; Row 2: controlled. Text under each panel shows perceived gender and race (A/B/W). Baseline (seed 0) DALL·E 3 Baseline (seed 0) FLUX.1-dev Baseline (seed 0) Gemini Imagen 4.0 Baseline (seed 0) Grok-2 Image Baseline (seed 0) Stable Diffusion XL Turbo Controlled (seed 0) Controlled (seed 0) Controlled (seed 0) Controlled (seed 0) Controlled (seed 0)… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative exemplars for CEO across five models. Row 1: baseline; Row 2: controlled. Captions under each panel indicate condition and seed. and SDXL flip from male to female athletes, while DALL·E 3 remains female and FLUX remains male. For the race attribute, it is o…
Figure 5
Figure 5. Figure 5: Gender composition across generative models. Controlled prompts generally increase [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Race composition across generative models (Hispanic excluded). Baseline generations [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 28 canonical work pages

  1. [1]

    Stereotyping and evaluation in implicit race bias: evidence for independent constructs and unique effects on behavior

    David M Amodio and Patricia G Devine. Stereotyping and evaluation in implicit race bias: evidence for independent constructs and unique effects on behavior. Journal of personality and social psychology, 91(4):652, 2006

    work page 2006

  2. [2]

    Training a helpful and harmless assistant with reinforcement learning from human feedback

    Yuntao Bai, Andy Jones, Kamal Ndousse, Amanda Askell, Anna Chen, Nova DasSarma, Dawn Drain, Stanislav Fort, Deep Ganguli, Tom Henighan, and others. Training a helpful and harmless assistant with reinforcement learning from human feedback. arXiv preprint arXiv:2204.05862, 2022

    arXiv 2022

  3. [3]

    Improving image generation with better captions

    James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, et al. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3):8, 2023. 8

    work page 2023

  4. [4]

    Gender shades: Intersectional accuracy disparities in com- mercial gender classification

    Joy Buolamwini and Timnit Gebru. Gender shades: Intersectional accuracy disparities in com- mercial gender classification. In Proceedings of the 1st Conference on Fairness, Accountability and Transparency (FAT*), volume 81 of Proceedings of Machine Learning Research, pages 77–91. PMLR, 2018

    work page 2018

  5. [5]

    Biasmap: Can cross-attention uncover hidden social biases? In Workshop on Demographic Diversity in Computer Vision@ CVPR 2025

    Rajatsubhra Chakraborty, Xujun Che, Depeng Xu, Cori Faklaris, Xi Niu, and Shuhan Yuan. Biasmap: Can cross-attention uncover hidden social biases? In Workshop on Demographic Diversity in Computer Vision@ CVPR 2025

    work page 2025

  6. [6]

    TIBET: Identifying and evaluating biases in text-to-image generative models

    Aditya Chinchure, Saifuddin Syed, Jean Kaddour, Luca Soldaini, Zaïd Harchaoui, et al. TIBET: Identifying and evaluating biases in text-to-image generative models. In Computer Vision – ECCV 2024, 2024. LNCS, Springer

    work page 2024

  7. [7]

    DALL-Eval: Probing the reasoning skills and social biases of text-to-image generation models

    Jaemin Cho, Abhay Zala, and Mohit Bansal. DALL-Eval: Probing the reasoning skills and social biases of text-to-image generation models. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023

    work page 2023

  8. [8]

    Precisedebias: An automatic prompt engi- neering approach for generative ai to mitigate image demographic biases

    Colton Clemmer, Junhua Ding, and Yunhe Feng. Precisedebias: An automatic prompt engi- neering approach for generative ai to mitigate image demographic biases. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 8596–8605, 2024

    work page 2024

Show all 36 references
  1. [9]

    Eagly and Steven J

    Alice H. Eagly and Steven J. Karau. Role congruity theory of prejudice toward female leaders. Psychological Review, 109(3):573–598, 2002

    2002

  2. [10]

    Auditing and instructing text-to-image generation models for fairness

    Florian Friedrich, Zekai Chen, Allan Hanbury, and Michael Wiegand. Auditing and instructing text-to-image generation models for fairness. AI and Ethics, 2024

    2024

  3. [11]

    Google gemini imagen 4: Ai image generation

    Google DeepMind. Google gemini imagen 4: Ai image generation. https://gemini. google/overview/image-generation/, 2025. Accessed: 2025-08-25

    2025

  4. [12]

    Madeline E. Heilman. Gender stereotypes and workplace bias. Research in Organizational Behavior, 32:113–135, 2012

    2012

  5. [13]

    Gender and racial bias in visual question answering datasets

    Yusuke Hirota, Yuta Nakashima, and Noa Garcia. Gender and racial bias in visual question answering datasets. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, pages 1280–1292, 2022

    2022

  6. [14]

    Social biases in nlp models as barriers for persons with disabilities

    Ben Hutchinson, Vinodkumar Prabhakaran, Emily Denton, Kellie Webster, Yu Zhong, and Stephen Denuyl. Social biases in nlp models as barriers for persons with disabilities. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5491–550...

    2020

  7. [15]

    Fairface: Face attribute dataset for balanced race, gender, and age for bias measurement and mitigation

    Kimmo Karkkainen and Jungseock Joo. Fairface: Face attribute dataset for balanced race, gender, and age for bias measurement and mitigation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 1548–1558, 2021

    2021

  8. [16]

    Black Forest Labs. Flux. https://github.com/black-forest-labs/flux , 2024

    2024

  9. [17]

    Aligning text-to-image models using human feedback

    Kimin Lee, Hao Liu, Moonkyung Ryu, Olivia Watkins, Yuqing Du, Craig Boutilier, Pieter Abbeel, Mohammad Ghavamzadeh, and Shixiang Shane Gu. Aligning text-to-image models using human feedback. arXiv preprint arXiv:2302.12192, 2023

    2023 arXiv

  10. [18]

    Holistic evaluation of text-to-image models

    Tony Lee, Michihiro Yasunaga, Chenlin Meng, Yifan Mai, Joon Sung Park, Agrim Gupta, Yunzhi Zhang, Deepak Narayanan, Hannah Teufel, Marco Bellagente, Minguk Kang, Taesung Park, Jure Leskovec, Jun-Yan Zhu, Fei-Fei Li, Jiajun Wu, Stefano Ermon, and Percy Liang. Holistic evaluatio...

    2023

  11. [19]

    Fair text-to-image diffusion via fair mapping

    Jia Li, Lijie Hu, Jingfeng Zhang, Tianhang Zheng, Hua Zhang, and Di Wang. Fair text-to-image diffusion via fair mapping. arXiv preprint arXiv:2311.17695, 2023

    2023 arXiv

  12. [20]

    Sport–gender stereotypes and their impact on impression evaluations

    Zhiyuan Liu, Menglu Shentu, Yuhan Xue, Yike Yin, Zhihao Wang, Liangchen Tang, Yu Zhang, and Weiqi Zheng. Sport–gender stereotypes and their impact on impression evaluations. Hu- manities and Social Sciences Communications, 10(1):1–14, 2023. 9

    2023

  13. [21]

    Bender, et al

    Alexandra Sasha Luccioni, Yacine Jernite, Rishi Bommasani, Marzyeh Ghassemi, Emily M. Bender, et al. Stable bias: Evaluating societal representations in diffusion models. In Advances in Neural Information Processing Systems (NeurIPS) Datasets and Benchmarks Track, 2023. arXiv:...

    2023 arXiv

  14. [22]

    Relevancy and Diversity in News Recommendations

    Shaina Raza and Chen Ding. Relevancy and Diversity in News Recommendations. SIGIR, News Recommendation and Analytics (INRA 2022), 2022

    2022

  15. [23]

    Who is responsible? the data, models, users or regulations? responsible generative ai for a sustainable future

    Shaina Raza, Rizwan Qureshi, Anam Zahid, Joseph Fioresi, Ferhat Sadak, Muhammad Saeed, Ranjan Sapkota, Aditya Jain, Anas Zafar, Muneeb Ul Hassan, et al. Who is responsible? the data, models, users or regulations? responsible generative ai for a sustainable future. arXiv prepri...

    2025

  16. [24]

    Ridgeway

    Cecilia L. Ridgeway. Framed by Gender: How Gender Inequality Persists in the Modern World. Oxford University Press, New York, 2011

    2011

  17. [25]

    Beyond content: How grammatical gender shapes visual representation in text-to-image models

    Muhammed Saeed, Shaina Raza, Ashmal Vayani, Muhammad Abdul-Mageed, Ali Emami, and Shady Shehata. Beyond content: How grammatical gender shapes visual representation in text-to-image models. arXiv preprint arXiv:2508.03199, 2025

    2025

  18. [26]

    Fairt2i: Mitigating social bias in text-to-image generation via large language model-assisted detection and attribute rebalancing

    Jinya Sakurai and Issei Sato. Fairt2i: Mitigating social bias in text-to-image generation via large language model-assisted detection and attribute rebalancing. arXiv preprint arXiv:2502.03826, 2025

    2025

  19. [27]

    LAION-5B: An open large-scale dataset for training next generation image-text models

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Florian Böttger, Andreas Köpf, Björn Ommer, et al. LAION-5B: An open large-scale dataset for training next generation image-text models. arXiv preprint arXiv:2210.08402, 2022

    2022 arXiv

  20. [28]

    The bias amplification paradox in text-to- image generation

    Preethi Seshadri, Sameer Singh, and Yanai Elazar. The bias amplification paradox in text-to- image generation. arXiv preprint arXiv:2308.00755, 2023

    2023 arXiv

  21. [29]

    Representation bias in data: A survey on identification and resolution techniques

    Nima Shahbazi, Yin Lin, Abolfazl Asudeh, and HV Jagadish. Representation bias in data: A survey on identification and resolution techniques. ACM Computing Surveys, 55(13s):1–39, 2023

    2023

  22. [30]

    Stable diffusion xl turbo

    Stability AI. Stable diffusion xl turbo. https://huggingface.co/stabilityai/ sdxl-turbo, 2023. Accessed: 2025-08-25

    2023

  23. [31]

    Smiling women pitching down: auditing representational and presentational gender biases in image-generative ai

    Luhang Sun, Mian Wei, Yibing Sun, Yoo Ji Suh, Liwei Shen, and Sijia Yang. Smiling women pitching down: auditing representational and presentational gender biases in image-generative ai. Journal of Computer-Mediated Communication, 29(1):zmad045, 2024

    2024

  24. [32]

    Assessing social and intersectional biases in contextualized word representations

    Yi Chern Tan and L Elisa Celis. Assessing social and intersectional biases in contextualized word representations. Advances in neural information processing systems, 32, 2019

    2019

  25. [33]

    Hartley, and Ajmal Mian

    Jordan Vice, Naveed Akhtar, Richard I. Hartley, and Ajmal Mian. Quantifying bias in text-to- image generative models. arXiv preprint arXiv:2312.13053, 2023

    2023 arXiv

  26. [34]

    Wenxuan Wang, Haonan Bai, Jen tse Huang, Yuxuan Wan, Youliang Yuan, Haoyi Qiu, Nanyun Peng, and Michael R. Lyu. New job, new gender? measuring the social bias in image generation models. arXiv preprint arXiv:2401.00763, 2024. ACM MM 2024 Oral

    2024 arXiv

  27. [35]

    Evaluating fairness in large vision-language models across diverse demographic attributes and prompts

    Xuyang Wu, Yuan Wang, Hsin-Tai Wu, Zhiqiang Tao, and Yi Fang. Evaluating fairness in large vision-language models across diverse demographic attributes and prompts. arXiv preprint arXiv:2406.17974, 2024

    2024

  28. [36]

    An athlete running in a stadium

    xAI. xai api guide: Image generations. https://docs.x.ai/docs/guides/ image-generations, 2025. Accessed: 2025-08-25. 10 Appendix A Prompts for Content (Image) Generation For transparency and reproducibility, we include the exact prompts used in our experiments, along with a co...

    2025

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.