LLMs show surprising strengths in some low-resource languages — but real challenges remain in cultural nuance, temporal reasoning, and ethical risks. We explore where AI succeeds, where it struggles, and what nonprofits need to know before adopting multilingual AI.