Artificial Intelligence (AI) is transforming the discipline of sociolinguistics by reshaping methodologies, introducing new modes of language production and influencing social interaction and power dynamics mediated by language. This paper explores the dual role of AI as both a research instrument and a linguistic agent in society, with a special emphasis on mental health discourse. Drawing on real-world case studies and interdisciplinary scholarship, we highlight how AI contributes to detecting linguistic markers of distress, mediating stigma, enhancing mental health communication and co-constructing supportive digital environments. We argue that AI-driven language technologies must be informed by sociolinguistic expertise to ensure equitable, inclusive and ethically responsible applications, particularly in sensitive domains like mental health.