Zero-shot learning is a machine learning method that allows pre-trained models to classify data based on labels that have not been used to train the model. However, creating labels for zero-shot classification can be time-consuming and error-prone. Using OpenAI’s GPT-3, a set of labels relevant to the data can be generated, significantly reducing the effort required to create labels. A function to generate relevant labels and to classify the data based on those labels is defined. The end-to-end zero-shot classifier automates the label generation process using GPT-3 and can be used for zero-shot text classification.