The ongoing and rapid development of artificial intelligence (AI) is transforming multilingual communication in education and healthcare, two fields where linguistic diversity directly affects access and equity. This article investigates how AI-based translation tools, Google Translate, DeepL and ChatGPT - 4, perform within the multilingual environment of Vasile Goldis Western University of Arad in Romania. Using a mixed-methods approach, the study combined quantitative evaluation of translation accuracy with qualitative interviews involving 60 participants, including students, faculty clinicians, and administrative staff. Findings indicate that while ChatGPT – 4 and DeepL achieve high levels of fluency and semantic accuracy, users remain cautious about trust, confidentiality, and cultural sensitivity, especially in high stakes contexts such as clinical documentation. Across both educational and healthcare domains, AI tools were valued for speed and convenience but criticised for their inability to convey nuance, empathy, and intercultural awareness. This article argues that translation accuracy, though necessary, is insufficient on its own. For AI to strengthen multilingual communication, institutions must prioritise linguistic equity, critical AI literacy, and ethical oversight, ensuring that technological gains serve support inclusive and context sensitive communication.
Stiegelbauer Laura-Rebeca (Tue,) studied this question.