This study explores the application of Generative Artificial Intelligence in the identification and validation of collocations, employing ChatGPT and Gemini tools with some of their respective models. By leveraging Few–Shot and Zero–Shot prompt engineering techniques, the research evaluates the model’s effectiveness in collocation analysis. Furthermore, the study highlights the importance of efficiently grouping collocations within prompts to enhance computational efficiency while maintaining classification performance. Although Artificial Intelligence (AI) presents significant potential in assisting linguistic research, human validation remains essential to mitigate false positives and negatives. Future studies may refine model training and explore hybrid methodologies integrating automation with human expertise to enhance reliability. The findings contribute to the growing field of AI–assisted linguistic analysis, underscoring the need for strategic application in academic and professional contexts.