The complexity of gender and language

Abstract

In this paper, we present a physical board game, known as the Gender Game1(GG), and detail how we enhanced it with digital elements. The goal of the GG is to deconstruct gender stereotypes by highlighting a scientific approach: the GG is rooted in social sciences and aims to inform the public about gender equality by presenting scientific results, mainly from research projects. The game is played with several players seated around a giant board; players take turns to roll the dice, pick a card from the deck held by the moderator and move their pawn to a domain where stereotypes exist. After a question is shown on a screen, the moderator standing in the middle asks the team players a question and all players discuss the answer together. A core component of the game addresses gendered language directly, serving as a practical tool for promoting linguistic rights and raising awareness of how language shapes perceptions of gender and power. We conducted an exploratory study with generative artificial intelligence (AI) by submitting some of the GG questions to ChatGPT 3.5 and 4.0. In this paper, we discuss the results of this study. Generative AI and, in particular, Large Language Models (LLMs) are impactful tools for decision-making and process automation. However, bias is often found in these tools and is perpetuated in the generated content. Thus, an ethical and responsive usage of LLMs is of utmost importance. From this perspective, the GG is positioned as a proactive, human-centred intervention designed to address the root cause of the biases that AI amplifies.

Ähnliche Arbeiten