Multilingual AI Story Generation System Using Natural Language Processing and Machine Translation

Abstract

The rapid evolution of artificial intelligence has significantly transformed the field of natural
language processing (NLP), enabling machines to generate human-like text with remarkable
accuracy. One of the most promising applications of NLP is automated story generation, which
has gained popularity in creative writing, education, and entertainment. This project presents a
multilingual AI story generation system capable of generating and translating stories into regional
Indian languages such as Telugu, Tamil, and Kannada.
The system utilizes a pre-trained transformer-based language model (GPT-2) for generating
coherent and contextually relevant stories from user-provided prompts. The generated stories are
initially created in English and then translated into the selected target language using a machine
translation model. This approach ensures both high-quality text generation and accessibility for
users who prefer regional languages.
A user-friendly graphical interface is developed using the Tkinter library, allowing users to input
story ideas, select languages, generate stories, expand narratives, and save outputs. The system
supports Unicode rendering to display text in native scripts such as Telugu (తెలుగు), Tamil
(தமிழ்), and Kannada (ಕನ್ನಡ), enhancing readability and user experience.
The system also includes an elaboration feature, which expands the generated story by adding
emotions, dialogues, and descriptive elements. This is achieved by re-feeding the generated story
into the model with an enhanced prompt, demonstrating the capability of iterative text refinement.
Performance evaluation is based on qualitative analysis, focusing on coherence, creativity, and
linguistic accuracy. While GPT-2 provides strong baseline performance, translation quality may
vary depending on language complexity and context.