AI-assisted practical guide. Examples are hypothetical; these are proposed editorial methods, not reported research results.
Generating reading‑comprehension exercises from fiction works best when the prompt guides the AI to produce three clearly distinct question types: explicit, inference and imagination. Explicit questions ask students to locate a word‑for‑word answer in the text; inference questions require them to combine two or more pieces of information to reach a logical conclusion; imagination questions invite them to extend the story in a way that is supported by the mood, setting or character traits but not predetermined by the passage. By separating these levels, the exercise moves beyond simple recall and encourages deeper analytical thinking.
Structuring the Prompt
Ask the AI to act as a curriculum designer and to analyze a supplied excerpt. The prompt should request three questions, each labeled with its category, and a brief justification for why the question belongs in that category. For example:
“Read the following passage and create one explicit question, one inference question, and one imagination question. After each question, include a one‑sentence explanation of the category it represents.”
Explicit questions must be answerable by quoting a single phrase from the text. Inference questions should draw on at least two textual clues and lead to a conclusion that is not stated verbatim. Imagination questions must be open‑ended, encouraging students to propose plausible continuations that are consistent with the established tone, setting or character motivations, without a single correct answer. Finally, add a step that a human reviewer checks each AI‑generated item to confirm that the categorization and answer key are accurate.
Hypothetical example
Suppose the AI receives the following short scene:
Elias paced the dimly lit room, glancing repeatedly at the silent telephone on the table. Each time the phone remained still, his shoulders slumped a little more.
A well‑crafted prompt could yield:
Explicit: What object does Elias check repeatedly? – The telephone. (The answer appears directly in the sentence.)
Inference: What emotion is most likely affecting Elias as he waits? – Anxiety, because his pacing and repeated checking suggest nervous anticipation. (The answer is inferred from two textual clues: pacing and repeated glances.)
Imagination: How might the story change if the telephone finally rang, and what are two different plausible reasons for the call? – Students could suggest a warning from a mysterious stranger, a long‑lost relative seeking reconciliation, or an unexpected job offer, each justified by the tense atmosphere. (Because the text does not specify who calls, multiple creative responses are possible.)
Each question is followed by a brief note explaining its classification, and a human reviewer would verify that the imagination question indeed allows several valid answers.
Verifying the Exercise
After the AI produces the set, a human reviewer should read the passage, the questions and the provided answer key. First, confirm that the explicit question’s answer is a verbatim excerpt; if any wording differs, revise the question. Next, ensure the inference question requires synthesis of at least two textual details and that its answer is not stated outright. Finally, check that the imagination question is open‑ended: it must invite more than one reasonable continuation and should not be answerable by a single fact from the text. If any question fails its test, rewrite it and re‑run the AI or adjust the prompt accordingly. This human check helps check whether the final exercise reliably distinguishes explicit knowledge, logical inference and creative imagination, fulfilling the intended learning goal.