Is using AI for homework cheating?

Almost every hard case resolves with two questions. Can you explain every part of what you handed in (not recite it, explain it, including why it is phrased that way)? And was anyone misled about the help you had?

Pass both and it is almost never cheating. Fail the first and you have handed in something you do not understand, which will find you out at the next test regardless of any rule. Fail the second and the problem is not the AI at all. It is the dishonesty, which would be just as serious if the help had come from an older sibling.

We teach it as those two questions rather than as a list of permitted tools, because the list goes out of date every few months and the questions do not.

"Used AI" covers behaviors with nothing in common. Asking it to quiz you before a test, to explain a concept a different way, or to tell you which part of your argument is weakest: those are study skills, and good ones. They leave the thinking with the student.

Further along: asking it to check your grammar, to suggest a structure for an essay you have already planned, or to summarize a source you have also read yourself. These are usually fine and are increasingly ordinary, but they are exactly where a school policy may draw its own line, so it is worth knowing what that line is.

At the far end: asking it to draft the paragraphs, to produce the analysis, or to write the thing and then changing some words. Here the student cannot explain what they handed in, and the two-part test fails on the first question before anyone even gets to honesty.

The useful conversation with a child is not "did you use AI" but "which of those was it?"

Policies differ enormously right now, and they are being rewritten constantly. One teacher may require AI use to be disclosed in a footnote; another may forbid it for a particular assignment; a third may build the assignment around it.

So the first responsibility we put on a student is to know the actual rule for the actual class, and to ask when it is not stated. "I did not know" is a weak position and an avoidable one. A student who asks a teacher in advance how AI may be used on an assignment has never once made things worse for themselves.

Plenty of students have already handed in something they could not fully explain, usually in a panic the night before. Treating it as a character failure tends to end the conversation and drive the next instance underground.

More useful: work out what the pressure actually was. Nearly always it was time, or not understanding the material and not wanting to say so. Both of those have better solutions, and AI is genuinely one of them. A student who cannot start an essay can ask for three possible angles and then argue with all three, and end up with more of their own thinking on the page than they would have had otherwise.