Developing AI Chatbots for Comedic Persona Preservation in Russian Translation: Kevin Hart (ESFP) and Louis C.K. (INFP)

Abstract

The fast-paced development of artificial intelligence has opened up new frontiers in the field of translation studies, but yet the conveyance of comedic discourse remains a key problem. Humor is a matter not just of language but of comedic timing, of the individual performer’s style, and of the societal expectations of the audience. The present work is devoted to the creation of individualized AI chatbots to keep the humorous identity when translating stand-up into Russian. The study investigates two conflicting types of comedy from the MBTI psycholinguistic typology: Kevin Hart’s expressive extraversion (ESFP) and Louis C. K.’s reflective self-irony (INFP). Using mixed methods, the authors compare the main language model (ChatGPT) with a custom chatbot (built on the Google Gemini API) trained through prompt engineering. The results indicate that the traditional machine tends to over-euphemize and lose pragmatics, while the personalized AI successfully adopts direct, adaptive and selective translation tactics. The bespoke model, addressing the challenge of non-normative vocabulary and culturally distinct realities, considering the local context of the Gen Z youth audience in Kazakhstan, statistically achieves the level of professional human translation. The research shows that the algorithmic use of psycholinguistic characteristics promotes the pragmatic equivalence and authenticity of translation, notwithstanding the observed restrictions concerning the possibility of pragmatic “over-localization.