Since its first release (2022), ChatGPT became part of the scientists’ workflow, both as tool and research subject. The awesome opportunities related to such powerful application increased the interests of scholars for Artificial Intelligence (AI) and Natural Language Processing (NLP), attracting not only computer scientists, AI and data architects and computational linguists, but also (digital) humanists, philosophers, and literary scholars. ChatGPT, and Large Language Models’ (LLM) application in text generation, however, are not the only possible NLP outcome in language and text sciences in the broadest sense. In this paper, we will present part of our current research program, which is focused on investigating some Italian language varieties with the help of AI and data science. Our goals are multiple: (a) testing the existence of some presumed obsolete varieties; (b) defining boundaries between similar linguistic varieties and, more in general, measuring the distance between language varieties; and (c) defining a set of language features and testing the weight of them in assigning a given utterance to its language variety. We will present the results of our first experiment, as well as our next scheduled research steps, and we will reflect on how AI can help sociolinguists and language historians in their daily work.