Large language models (LLMs) have been prominent in various tasks, including text generation and summarisation. The applicability of LLMs to the generation of product reviews is gaining momentum, paving the way for the generation of movie reviews. In this study, we propose a framework that generates movie reviews using three LLMs (GPT-4o, DeepSeek-V3, and Gemini-2.0), and evaluate their performance by comparing the generated outputs with IMDb user reviews. We use movie subtitles and screenplays as input to the LLMs and investigate how they affect the quality of reviews generated. We review the LLM-based movie reviews in terms of vocabulary, sentiment polarity, similarity, and thematic consistency in comparison to IMDB user reviews. The results demonstrate that LLMs are capable of generating syntactically fluent and structurally complete movie reviews. Nevertheless, there is still a noticeable gap in emotional richness and stylistic coherence between LLM-generated and IMDb reviews, suggesting that further refinement is needed to improve the overall quality of movie review generation. We provided a survey-based analysis where participants were told to distinguish between LLM and IMDb user reviews. The results show that LLM-generated reviews are difficult to distinguish from IMDB user reviews. We found that DeepSeek-V3 produced the most balanced reviews, closely matching IMDb reviews. GPT-4o overemphasised positive emotions, while Gemini-2.0 captured negative emotions better but showed excessive emotional intensity.

Sentiment analysis, also called opinion mining, is a typical application of Natural Language Processing (NLP) widely used to analyze a given sentence or statement's overall effect and underlying sentiment. A sentiment analysis model classifies the text into positive or negative (and sometimes neutral) sentiments in its most basic form. Therefore naturally, the most successful approaches are using supervised models that need a fair amount of labelled data to be trained. Providing such data is an expensive and time-consuming process that is not possible or readily accessible in many cases. Additionally, the output of such models is a number implying how similar the text is to the positive examples we provided during the training and does not consider nuances such as sentiment complexity of the text.