Digital games have relied on forms of artificial intelligence (AI) since the industry began in the 1970s. The AI used in early games, like those made for the Atari 2600, primarily drew on rule-based scripts to control the behaviour of its nonplayer characters (NPCs) (Yannakakis and Togelius, Artificial intelligence and games. Springer, 2018). The AI technology of today, however, can beat all 57 original Atari games ‘with performance above the human benchmark’ (Badia, Agent57: outperforming the Atari human benchmark. In: Proceedings of the 37th International Conference on Machine Learning, vol 119, pp 507–517. Available from https://proceedings.mlr.press/v119/badia20a.html , 2020, p. 8). While AI has a long history with digital games designed for entertainment, the use of AI in educational apps or games designed for second-language (L2) learning is garnering great interest from educators, researchers and language learners. This interest has been driven by the now widely available large language models (LLMs) that power generative AI platforms like ChatGPT (Open AI, ChatGPT [Large language model application]. https://chat.openai.com/ , 2022) and Claude (Anthropic, Claude [Large language model]. https://claude.ai/ , 2023), among others. Generative AI can process and generate language that is remarkably similar to human language, leading to exciting new areas of investigation in the domain of Digital Game Based Language Learning and Teaching (DGBLLT). In this entry, the evolution of AI in digital games is summarized as well as the technology driving generative AI. Findings and implications from this new area of research that investigates its use in the DGBLLT domain are also reviewed. This review is followed by a discussion of the ethics and risks related to generative AI that stem from the inherent bias in the language data with which the LLMs are trained. We conclude with direction for future research on incorporating generative AI into L2 classrooms and DGBLLT contexts.