Gender Biases in Generative AI: Unveiling Prejudices and Prospects in the Age of ChatGPT

Abstract

The advent of Advanced Natural Language Processing Models and generative AI, exemplified as ChatGPT, has exerted a significant impact on individuals' personal lives and professional endeavors since its introduction, with more expansion anticipated in the future. This study delves into the intricate landscape of gender biases entrenched within Generative Artificial Intelligence (AI), specifically focusing on the prevalence and implications within the domain of ChatGPT. Through an extensive exploration of ChatGPT's responses and interactions, this research sheds light on the nuanced manifestations of gender stereotypes, disparities, and their multifaceted impact on societal constructs. Highlighting instances of bias across various prompts, including professional scenarios, parental anecdotes, and skill prioritization in CVs, the study delineates the perpetuation of gendered notions in AI-generated content. Moreover, it underscores the perlocutionary effects of such biases, elucidating their potential to reinforce societal disparities. The research underscores the significance of ethical frameworks and regulatory measures to counteract these biases, emphasizing the pivotal role of AI in promoting transformative change and fostering gender equality. Ultimately, this inquiry advocates for an informed and proactive approach to harness the promising potential of AI while mitigating gender prejudices in the digital age. Nevertheless, this study also proposes that artificial intelligence has the potential to address prejudices and counteract gender disparities. The present discourse is around the topic of gender prejudice in the context of ChatGPT, a prominent example of big language models utilized in generative AI. The focus lies on the examination of performativity and ethical considerations associated with AI systems.