. // Multilingual gender fairness in AI-generated texts

Topic description

.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Despite major advances in GenAI, stereotyping and biases are ubiquitous in texts generated by AI technologies, which runs counter to a European society committed to gender equality. The doctoral contract will:

1) investigate texts generated by leading LLMs (e.g. OpenAI's ChatGPT, Google's Gemini, Mistral's Le Chat) in terms of gender fairness with a focus on languages other than EN (e.g. DA, FR, DE);

2) explore the impact of linguistic factors (i.e. linguistic bias), specifically the effects of grammatically un/gendered languages and the masculine default, on gendered representat...