What it does well, and where it fails

Home / AI literacy guides / AI basics

AI literacy guide · AI basics

What it does well, and where it fails

Match the tool to tasks it is good at, and know where it breaks.

Generative AI is genuinely useful for some tasks: drafting and rephrasing text, summarizing, brainstorming options, explaining a concept several ways, and generating examples or practice items. For a busy instructor, those are real time savers.

It fails in predictable ways. It fabricates facts, sources, and quotations, a behavior often called hallucination, and states them confidently. It is weak on events past its training, on precise computation, and on anything needing genuine reasoning or lived experience. It does not know your students, your course, or your context. The practical rule is to use it where a fluent first draft helps and a human will verify, and to avoid it where accuracy you cannot check is essential.