In recent years machine translation (MT) has been engaged more frequently both in foreign language (FL) classes and in our daily lives. Although instructors often remain concerned about students using MT for FL learning, recent studies show that the trend is inevitable and therefore instructors need to find ways to adapt to MT. Current studies show that MT can make contributions in various ways but students at different levels have different focal points. This study uses multiple case studies to find out how students at different proficiency levels use MT, how they make decisions on what to revise and why they make such decisions. The study was conducted with five participants of students studying Japanese at Western universities. The study analyses how participants write in Japanese and make edits using Google translate (GT) as a MT tool; the study also verified some of the previous findings in the Japanese learning environment and identified several distinctive characteristics of students using GT in Japanese learning.