CoChat
Sign Up
Cited by 5Ethics and Social Impacts of AIHate Speech and Cyberbullying DetectionMedia, Gender, and Advertising

Understanding Gender Bias in AI-Generated Product Descriptions

Open access
Kelly, Markelle, Tahaei, Mohammad, Smyth, Padhraic, Wilcox, Lauren·2025·Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency
DOI 10.1145/3715275.3732169
Open this paper in CoChat — free accountOpen PDF

Abstract

While gender bias in large language models (LLMs) has been extensively studied in many domains, uses of LLMs in e-commerce remain largely unexamined and may reveal novel forms of algorithmic bias and harm.Our work investigates this space, developing data-driven taxonomic categories of gender bias in the context of product description generation, which we situate with respect to existing general purpose harms taxonomies.We illustrate how AI-generated product descriptions can uniquely surface gender biases in ways that require specialized detection and mitigation approaches.Further, we quantitatively analyze issues corresponding to our taxonomic categories in two models used for this task-GPT-3.5 and an e-commerce-specific LLM-demonstrating that these forms of bias commonly occur in practice.Our results illuminate unique, under-explored dimensions of gender bias, such as assumptions about clothing size, stereotypical bias in which features of a product are advertised, and differences in the use of persuasive language.These insights contribute to our understanding of three types of AI harms identified by current frameworks: exclusionary norms, stereotyping, and performance disparities, particularly for the context of e-commerce.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1145/3715275.3732169

Cites

94 of 94 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 19

  1. Leon Derczynski, Hannah Rose Kirk, Vidhisha Balachandran, Sachin Kumar, Yu- lia Tsvetkov, and Saif Mohammad. 2023. Assessing language model deployment with risk cards. arXiv preprint arXiv:2303.18190 (2023). · arXiv
  2. Seliem El-Sayed, Canfer Akbulut, Amanda McCroskery, Geoff Keeling, Zachary Kenton, Zaria Jalan, Nahema Marchal, Arianna Manzini, Toby Shevlane, Shan- non Vallor, et al. 2024. A mechanism-based approach to mitigating harms from persuasive generative AI. arXiv preprint arXiv:2404.15058 (2024). · arXiv
  3. F Gasparini, I Erba, E Fersini, S Corchs, et al. 2018. Multimodal classification of sexist advertisements. In ICETE 2018-Proceedings of the 15th International Joint Conference on e-Business and Telecommunications-Volume 2, Vol. 1. SciTePress, 399-406.
  4. Mahammed Kamruzzaman, Hieu Minh Nguyen, and Gene Louis Kim. 2024. "Global is good, local is bad?": Understanding brand bias in LLMs. arXiv:2406.13997 [cs.CL] https://arxiv.org/abs/2406.13997 · arXiv
  5. Haein Kong, Yongsu Ahn, Sangyub Lee, and Yunho Maeng. 2024. Gender bias in LLM-generated interview responses. In Workshop on Socially Responsible Language Modelling Research.
  6. Fajri Koto, Jey Han Lau, and Timothy Baldwin. 2022. Can pretrained language models generate persuasive, faithful, and informative ad text for product descrip- tions?. In Proceedings of the Fifth Workshop on e-Commerce and NLP (ECNLP 5),
  7. Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michi- hiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, et al. 2023. Holistic evaluation of language models. Transactions on Machine Learning Research (2023).
  8. Liam Magee, Lida Ghahremanlou, Karen Soldatic, and Shanthi Robertson. 2021. Intersectional bias in causal language models. arXiv preprint arXiv:2107.07691 (2021). · arXiv
  9. Chester Palen-Michel, Ruixiang Wang, Yipeng Zhang, David Yu, Canran Xu, and Zhe Wu. 2024. Investigating LLM applications in e-commerce. arXiv preprint arXiv:2408.12779 (2024). · arXiv
  10. Anandi Ramamurthy and Kalpana Wilson. 2013. Racism, appropriation and resistance in advertising. Colonial Advertising & Commodity Racism 69, 4 (2013), 2.
  11. Rabia Rauf, Sohail Kamran, and Najeeb Ullah. 2019. Marketing Of skin fairness creams And consumer vulnerability. CITY UNIVERSITY RESEARCH JOURNAL 9, 3 (Oct. 2019). https://www.cusitjournals.com/index.php/CURJ/article/view/262
  12. Maribeth Rauh, John Mellor, Jonathan Uesato, Po-Sen Huang, Johannes Welbl, Laura Weidinger, Sumanth Dathathri, Amelia Glaese, Geoffrey Irving, Iason Gabriel, et al. 2022. Characteristics of harmful text: Towards rigorous bench- marking of language models. Advances in Neural Information Processing Systems 35 (2022), 24720-24739.
  13. Alexander Rogiers, Sander Noels, Maarten Buyl, and Tijl De Bie. 2024. Persuasion with Large Language Models: a Survey. arXiv preprint arXiv:2411.06837 (2024). · arXiv
  14. Alexandra A. Siegel. 2020. Online Hate Speech. Cambridge University Press, 56-88.
  15. Irene Solaiman, Zeerak Talat, William Agnew, Lama Ahmad, Dylan Baker, Su Lin Blodgett, Canyu Chen, Hal Daumé III, Jesse Dodge, Isabella Duan, et al. 2023. Evaluating the social impact of generative AI systems in systems and society. arXiv preprint arXiv:2306.05949 (2023). · arXiv
  16. Mohammad Tahaei, Marios Constantinides, Daniele Quercia, and Michael Muller. 2023. A Systematic Literature Review of Human-Centered, Ethical, and Respon- sible AI. arXiv:2302.05284 [cs.HC] https://arxiv.org/abs/2302.05284 · arXiv
  17. CWJ van Miltenburg. 2016. Stereotyping and bias in the flickr30k dataset. In 11th workshop on multimodal corpora: computer vision and language processing.
  18. Akshaj Kumar Veldanda, Fabian Grob, Shailja Thakur, Hammond Pearce, Ben- jamin Tan, Ramesh Karri, and Siddharth Garg. 2023. Are Emily and Greg still more employable than Lakisha and Jamal? Investigating algorithmic hiring bias in the era of ChatGPT. arXiv:2310.05135 [cs.CL] https://arxiv.org/abs/2310.05135 · arXiv
  19. Li Zhou, Taelin Karidi, Wanlong Liu, Nicolas Garneau, Yong Cao, Wenyu Chen, Haizhou Li, and Daniel Hershcovich. 2024. Does mapo tofu contain coffee? prob- ing LLMs for food-related cultural knowledge. arXiv preprint arXiv:2404.06833 (2024). · arXiv

Cited by

Sorted by citations

Related

Authors
Kelly, M., Tahaei, M., Smyth, P., & Wilcox, L.
Affiliations
  • University of California, Irvine (US)
  • eBay (United States) (US)
  • Georgia Institute of Technology (US)
ORCID
Venue
Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency
Publisher
ACM
Published
2025-06-23
Pages
2587-2615
DOI
10.1145/3715275.3732169
OpenAlex
W4411541874
Cited by
4
References
94
Citations by year
2026: 4
FWCI
4.06
Type
conference paper
Language
en
Open access
Licence
cc-by-nc
Locations
  • ACM Conference on Fairness, Accountability, and Transparency (FAccT) (publishedVersion, cc-by-nc)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
  • Social Sciences › Gender Studies › Media, Gender, and Advertising
Keywords
gender bias, large language models, e-commerce, algorithmic bias, stereotyping, persuasive language, GPT-3.5, performance gap

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review