Case study: translation from English into European Portuguese using gender-neutral language. Do AI chatbots perform better than MT Systems?

Abstract

There is a growing demand for gender-neutral writing in translation tasks motivated by the need for inclusion, neutrality or economic factors. Studies suggest that current Machine Translation (MT) and Artificial Intelligence (AI) systems tend to mirror the source text's (and real-world) gender bias or default generic masculine. It is important to understand whether, with documents in a gender-neutral style, the MT and AI tools provide better gender-neutral translations by resorting to standard and established grammatical solutions. Additionally, the effectiveness of AI systems, with customised prompts, in surpassing MT limitations regarding gender-neutral translation is also analysed.