Gender linguistics and artificial intelligence: Representation of femininity and masculinity (Based on invitations and congratulations)

Currently, artificial intelligence (AI) is actively being introduced into all scientific fields, gender linguistics being no exception. This study aims to compare texts produced by people and texts generated by neural network in 20 order to identify language means associated with gender. The material for the study included verbal invitations and congratulations in Russian and English created in accordance with teacher’s assignment by 25 young males and females aged 19–20 and the texts on similar issues generated by a neural network (ChatGPT). It was found that, first, females on average construct longer sentences and longer messages in general. Secondly, emoji and exclamatory marks were more abundant in female texts. Third, on the lexical level females use more emotive words and words with vague meanings but, in general, adhere to neutral style. At the same time, males more often use contractions and colloquial vocabulary. Finally, it was discovered that despite generally successful attempts of the neural network to apply verbal gender stereotypes, the texts generated by it display the difference with the texts produced by our male and female participants, showing schematism and sameness.