Understanding Gender Bias in AI-Generated Product Descriptions
This addresses algorithmic bias in e-commerce applications of LLMs, which is an under-explored domain with potential for novel harms.
The paper investigated gender bias in AI-generated product descriptions for e-commerce, developing a taxonomy of bias types and quantitatively demonstrating their prevalence in GPT-3.5 and an e-commerce-specific LLM, revealing issues like assumptions about clothing size and stereotypical feature advertising.
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.