Automating Sexual Injustice: Epistemic Injustice in Fembot Design and Feminist Directions for Equitable HRI

TL;DR

This paper critiques the design of female sex robots as biased and unjust, proposing feminist and evidence-based design directions to promote epistemic justice and inclusivity in HRI.

Contribution

It introduces feminist design principles grounded in epistemic justice to guide the development of equitable and inclusive sexual AI robots.

Findings

Analysis of fembots reveals male-centric biases and epistemic injustice.

Proposes three feminist design directions: empirical grounding, epistemic plurality, active consent.

Concrete evaluation criteria are provided for inclusive and just fembot design.

Abstract

Current AI-enabled female sex robots, or "fembots," are primarily designed to simulate female sexual responses through a lens of male-centric bias and pornographic stereotypes. This paper analyses fembot development as a failure in equitable robotics, arguing that these machines perpetuate "epistemic injustice" by prioritizing male hedonistic fantasies over empirical truths of female sexual experience in their design decisions. Drawing on Miranda Fricker's framework of testimonial and hermeneutical injustice, this analysis demonstrates how fembot interfaces discredit women's lived sexual knowledge and empirical research on female sexual physiology while privileging male-centred fantasies. This paper proposes three Feminist Design Directions including empirical grounding, epistemic plurality, and active consent modelling, which are grounded in Donna Haraway's concept of "Situated Knowledge" and accompanied by concrete evaluation criteria. These directions aim to facilitate a transition toward evidence-based intimate AI that prioritizes epistemic justice, mutuality, and inclusive design for marginalized users including disabled, neurodivergent, and LGBTQ+ communities.