Adultification Bias in LLMs and Text-to-Image Models

TL;DR

This paper investigates adultification bias in language and image generative AI models, revealing persistent biases against Black girls and highlighting the inadequacy of current alignment techniques to mitigate such biases.

Contribution

It introduces a new measurement of adultification bias in generative models and systematically analyzes this bias across text and image modalities.

Findings

LLMs show explicit and implicit adultification bias against Black girls.

T2I models depict Black girls as older and more revealingly dressed.

Current alignment methods are insufficient to fully address adultification bias.

Abstract

The rapid adoption of generative AI models in domains such as education, policing, and social media raises significant concerns about potential bias and safety issues, particularly along protected attributes, such as race and gender, and when interacting with minors. Given the urgency of facilitating safe interactions with AI systems, we study bias along axes of race and gender in young girls. More specifically, we focus on "adultification bias," a phenomenon in which Black girls are presumed to be more defiant, sexually intimate, and culpable than their White peers. Advances in alignment techniques show promise towards mitigating biases but vary in their coverage and effectiveness across models and bias types. Therefore, we measure explicit and implicit adultification bias in widely used LLMs and text-to-image (T2I) models, such as OpenAI, Meta, and Stability AI models. We find that LLMs exhibit explicit and implicit adultification bias against Black girls, assigning them harsher, more sexualized consequences in comparison to their White peers. Additionally, we find that T2I models depict Black girls as older and wearing more revealing clothing than their White counterparts, illustrating how adultification bias persists across modalities. We make three key contributions: (1) we measure a new form of bias in generative AI models, (2) we systematically study adultification bias across modalities, and (3) our findings emphasize that current alignment methods are insufficient for comprehensively addressing bias. Therefore, new alignment methods that address biases such as adultification are needed to ensure safe and equitable AI deployment.