Dr. Román Mendoza, Professor of Spanish at George Mason University, delivered a plenary talk titled "Possibilities and Limitations of Generative Artificial Intelligence for the Personalization of ELE/L2/LH Teaching" at the III Conference of Spanish Language Learning in Italy.
Her presentation opened by outlining the general limitations of AI and reviewing recent data on the market penetration of major chatbots, including ChatGPT, Claude, Gemini, DeepSeek, Copilot, and Grok, before examining how university students are currently using these tools for academic work. Turning to second-language classrooms, the talk compared human language with the synthetic language produced by large language models to assess how effectively chatbots can personalize the learning of a foreign, second, or heritage language, with Spanish as the primary case.
Grounded in both pragmatic and humanistic approaches to personalized learning, the presentation acknowledged that contemporary chatbots allow for a limited degree of customization through prompt engineering, retrieval-augmented generation (RAG), and the development of local LLMs. Nevertheless, Dr. Mendoza argued that inherent limitations of generative AI, including its inability to admit ignorance, reliance on biased or incomplete training data, and the fundamentally synthetic nature of its language, confine personalization largely to remedial functions, practice activities, and supplementary explanations. The talk concluded by noting that the absence of genuine communicative intent and a true hermeneutic contract significantly limits the value of chatbots as tools for fully personalized learning, particularly without teacher supervision.