This chapter discusses the impact of the recent advancement of machine translation (MT) engines on Japanese-as-a-foreign-language courses in higher education. It is argued that the communication-based approaches to foreign language pedagogy emphasizing information exchange will inevitably lose attractiveness because the real-time translation of text and speech by machine is becoming ubiquitously available. Instead, we propose a metacognitive approach, which foregrounds how we think rather than how we communicate. Shifting the curriculum focus from a communication-based to a thinking-based design will accentuate linguistic and cultural differences between learners’ L1 and L2, and such awareness likely inspires them to delve into their cognitive routines. This highly intellectual practice will assuredly preserve the vital relevance of foreign languages in liberal arts education. This chapter also demonstrates our experiment with ChatGPT and considers how its use can be incorporated into the Japanese language curriculum. It suggests that ChatGPT can make self-study of Japanese as natural as learning with a human teacher. However, at present, such a learning method is more suitable for intermediate to advanced learners than beginners.