Past, present, and future walked into a bar. It was tense. 🤪
This chapter covers AI prompts for grammar work: generating the six most common exercise types, producing learner-appropriate explanations without grammatical jargon, and using learners' native language to clarify concepts that are hard to explain in the target language alone.
|
What this chapter covers
|
How Do You Prompt AI to Create Grammar Exercises?
AI can generate common grammar exercise types with targeted prompts, e,g, gap-fill, error correction, transformation, contextualized practice, multiple choice, and word class identification — each suited to a different learning objective.
Grammar exercises are a staple of language teaching and one of the more repetitive parts of lesson prep. If you teach multiple students at different levels, you may find yourself creating variations of the same exercise types over and over. AI can help generate these quickly, and in this chapter we cover several common exercise types and how to prompt for them. Similar to the vocabulary prompts, being specific about the grammar point, the level, and the context helps the AI produce something useful. "Past tenses" is broad while "present perfect vs. past simple" is more specific and gives better results.
Gap-fill
Gap-fill (or cloze) exercises are the most common grammar exercise type. The learner fills in the missing word or form in a sentence:
Create a gap-fill exercise with 8 sentences for a B1-level adult ESL learner practicing the present perfect tense. Focus on recently completed actions affecting the present (e.g., "I have just finished..."). Use a context of daily routines and work life. Learner version: sentences with blanks. Teacher version: completed sentences with answers in bold.
The instruction about the specific use of present perfect ("recently completed actions affecting the present") helps narrow the exercise. Without it, AI may mix different uses of present perfect in ways that confuse learners at that level.
You can also provide the base verb for each blank, which makes the exercise about forming the correct tense rather than choosing the right word:
Create 8 gap-fill sentences for B1 learners practicing past simple vs. present perfect. Provide the base verb in brackets after each blank, e.g.: "She _______ (visit) Paris three times." Include an answer key with brief notes on why the correct tense is used.
Error correction
Error correction exercises give learners sentences with deliberate mistakes to find and fix:
Create an error correction exercise with 8 sentences for a B2-level learner. Each sentence should contain one grammar error related to conditional sentences (second and third conditional). Format: - The incorrect sentence - A hint about what type of error to look for - The corrected sentence (in a separate answer key)
Note that AI sometimes produces "errors" that are actually acceptable in informal speech or certain dialects. When reviewing error correction exercises, check that errors are genuinely wrong and not just a stylistic choice. If a sentence could reasonably be correct, it's not a good error correction item.
Transformation
Transformation exercises ask learners to rewrite sentences using a different structure while keeping the meaning:
Create a sentence transformation exercise with 6 items for B2-level learners. Each item should require transforming active voice to passive voice. Format each item as: Original: "The company launched a new product last month." Rewrite using passive voice: _______ Answer: "A new product was launched (by the company) last month."
Other common transformation types you can prompt for: direct speech to reported speech, positive to negative, combining two sentences with a relative clause, or rewriting using a specific word (key word transformation).
Contextualized practice
Rather than isolated sentences, you can ask for grammar embedded in a realistic scenario:
Create a short email exchange (3-4 emails) between two colleagues discussing a project. The emails should naturally use the first and second conditional throughout. After the emails, create 5 comprehension questions that require the learner to identify and explain the conditional forms used. Target level: B2.
This gives learners grammar practice within a context they might actually encounter, rather than a list of disconnected sentences.
Multiple choice
Create a multiple-choice grammar quiz with 8 questions for a B1-level learner. Topic: articles (a, an, the, zero article). Each question should be a sentence with a blank, followed by 4 options. The distractors should be plausible — focus on cases where article usage is commonly confused. Include an answer key.
Articles are a good example of where "plausible distractors" matters — in many cases more than one article could seem correct, and the exercise should test genuine understanding, not guessing.
Word class identification
You can also create exercises where learners identify grammar in existing text:
Write a short paragraph (80-100 words) about a weekend trip, using a mix of past simple, past continuous, and past perfect. Then create an exercise where the learner identifies and underlines:
- All past simple verbs (in one color)
- All past continuous verbs (in another color)
- All past perfect verbs (in a third color) Include a teacher answer key listing each verb and its tense.
As with vocabulary exercises, some AI chatbots offer a feature often called "canvas" or similar, which can generate interactive web pages you can share with a link. Grammar exercises like gap-fill, error correction, and multiple choice work well in this format. You can ask for this by adding something like "Create this as an interactive exercise where the learner fills in answers on screen and gets immediate feedback" to your prompt.
The different exercise types don't have to stand alone. You can prompt AI to generate a short sequence on the same grammar point: start with a gap-fill for recognition, follow with error correction for accuracy, then a contextualized dialogue for production. Try: something like the following.
Generate a 3-part exercise sequence for B1 learners on present perfect: first a gap-fill, then an error correction, then a short email they write using the present perfect.
How Do You Use AI to Explain Grammar to Learners?
AI can generate grammar explanations calibrated to any CEFR level, in plain language without metalinguistic jargon. This can save time when you need a clear, written explanation to share alongside exercises.
Explain the difference between the present perfect and the past simple in English. Write the explanation for a B1-level adult learner. Use simple language, short sentences, and 3-4 example sentences showing the contrast. Avoid grammatical terminology where possible — explain through examples.
The instruction to "avoid grammatical terminology where possible" helps produce explanations that learners can actually understand. If your learner is at a higher level or has studied linguistics, you can remove that constraint. You may also explicitly ask to make simpler explanations for a specific level.
Explain how English forms questions with auxiliary verbs (do/does/did). Write the explanation for an A2-level learner. Use only simple vocabulary. Give 3 examples and keep it under 150 words.
You can also ask for explanations and exercises to address common confusions your learner may be doing.
My B1 student keeps confusing "since" and "for" with the present perfect. Write a short explanation (5-6 sentences) with 4 example sentences that clearly show when to use each. Include 2 sentences where students commonly make mistakes.
Finally, you may also use AI to explain and illustrate exceptions to the rules.
Explain the main exceptions to the rule for forming the English past simple (irregular verbs). Write for a B1 learner. Group the irregular verbs by pattern where possible, and give 3 example sentences.
How Can AI Use Learners' Native Language for Grammar Teaching?
Grammar is one of the areas where L1 support is most effective. Concepts that are difficult to explain in the target language can be clarified quickly by comparing them to how the learner's own language handles the same idea. You may try to examine what differences exist compared to the learners first language with a prompt like the following.
My student is a Turkish-speaking B1 learner of English. We are working on the present perfect tense next week. Explain what aspects of the present perfect will be particularly difficult for a Turkish speaker, and why. Briefly describe how Turkish handles similar concepts. Then suggest 3 specific errors I should watch for.
You can use the result to inform your teaching or take it further and ask AI to generate exercises targeting those specific errors. To explain grammar in the learner's first language try a prompt like below, adjusted to your language combination.
Explain the English present perfect tense in Portuguese. Write it as if speaking to a native Portuguese speaker who is learning English. Compare the present perfect to how Portuguese expresses similar ideas. Include 5 example sentences with Portuguese translations.
If you don't speak the learner's L1, the caveats from Chapter 3 apply, so be transparent and let the learner verify.
What Should You Watch Out for When Using AI for Grammar?
Verify answer keys carefully. Grammar exercises are where answer key errors matter most. A wrong answer in the key can actively teach incorrect grammar. Always check, especially for exercises involving tenses where multiple answers might be debatable.
Be specific about the grammar point. "Past tenses" can mean past simple, past continuous, past perfect, or all three. "Conditionals" can mean zero, first, second, or third. The more specific you are, the more focused the exercise.
Watch for over-correction in error exercises. AI sometimes treats informal-but-correct usage as errors. "I've got a meeting" is perfectly acceptable, but AI may flag it in favor of "I have a meeting." Review error correction exercises to make sure the "errors" are genuinely wrong.
CEFR level accuracy varies. As with vocabulary, AI's sense of grammatical complexity is approximate. It may include structures that are above or below the requested level. A quick review helps catch these.
Request explanations in the answer key. Adding "with brief explanations of why the answer is correct" to your exercise prompts gives you material you can use when reviewing the exercise with your learner.