The Gendered, Epistemic Injustices of Generative AI

The rise of generative artificial intelligence (GenAI) brings optimism for productivity, economic, and social progress, but also raises concerns about algorithmic bias and discrimination. Regulators and theorists face the urgent task of identifying potential harms and mitigating risks. This article applies Miranda Fricker’s concepts of testimonial and hermeneutical injustice to explore how GenAI exacerbates or creates epistemic injustice from a feminist, epistemic perspective.Through three case studies, we reveal how gender-biased GenAI responses in leadership and workplace contexts reinforce stereotypes, leading to offline injustices. Moreover, we highlight how gender data gaps within the GenAI ecosystem contribute to hermeneutical injustice, marginalizing women’s experiences within collective knowledge resources. These findings demonstrate how GenAI technologies perpetuate both testimonial and hermeneutical injustices.By integrating existing work on epistemology, AI ethics, and feminist theory, we propose a novel framework for understanding the risks and harms of GenAI. We argue that meta-blindness further obscures these gendered issues and their potential solutions. Our analysis not only sheds light on these critical challenges but also offers a pathway toward achieving gender equity within the GenAI landscape.