Skip to main content

Building Prompts and Building Language: 3 Activities to Practice English While Building Prompts for ChatGPT

by Brent Warner |

Whether you’re looking at ChatGPT, Microsoft Bing, Google Bard, or any of the myriad chatbots that are flooding your social media accounts these days, you’re sure to have wondered about how your students might be using these tools in your class. There’s endless hand wringing about plagiarism (especially for those trying to define plagiarism in the AI era), but when we step away from the fretting and the fussing, we have an unprecedented opportunity to help our students build their language skills with what is now a free and widely available technology.

In a recent webinar from English Language Programs, Christina Cavage brought up the much overlooked point that  “prompt generation is language teaching, so teaching learners how to write prompts is teaching language.” This couldn’t be a more important concept to understand as students can have so much fun playing with these tools and learning at the same time.

When we stop looking at chatbots as solely output machines, and instead start to recognize that the output varies based on the quality of the input, it starts to become more and more clear that teaching students to compose high-quality messages for the bots will increase the fun they can have and also build their language competency. Let’s look at three ideas students can play with to build prompts that let them use their English to improve their English.

Beginning - Building Sentence Stems

When dealing with students with beginning English proficiency, one tried and true way to get them to build an intrinsic sense of English language patterns is through sentence stems. With a chatbot, they can start building out their own examples around topics of interest and play with them as much as they want.

Procedure

Introduction to Sentence Stems

Begin by introducing the concept of sentence stems. Explain that they are incomplete sentences where the learner fills in the blank to complete the thought.

Provide a few basic examples:

    • "I feel _____ when I _____."
    • "On weekends, I usually _____."

Brainstorming Session With Chatbots

Ask learners to brainstorm general topics or themes they're interested in (e.g., sports, music, holidays, family).

Using a chatbot, learners can ask for sentence stems related to their chosen topic. For instance, if a student chooses "sports," they might ask, "Can you give me sentence stems about sports?"

The chatbot can provide stems like:

    • "My favorite sport is _____ because _____."
    • "When I watch _____, I feel _____."

Follow-Up

While students are getting the results from the chatbot, you can write your own stems on the board:

“Can you make them ______?”

Students may find that the language produced by the chatbot is beyond their understanding, so this is a good opportunity to have them work with the chatbot to simplify the language. You can elicit adjectives from the class like “easier” or even go a little off track with things like “funnier” to see what it comes up with!

“Please give ____ examples for me to practice with, then check my ________”

Again, when students look up, you can discuss this stem and elicit ideas on how much practice students might want, and what specifically they want to focus on. Some students might say 5 examples and want it to check their spelling, while others might ask for 100 and want it to check their vocabulary.

Peer Review and Practice

At this point, students will be heavily involved with simple prompting, but now it’s time to mix things up:

    • Pair learners and have them exchange their chatbot prompts.
    • Each learner should respond to their partner's prompts, practicing their English and providing feedback to their classmates on what they might change.
    • Finally, come back to the class and reflect with everyone on their prompts and how they can use them in the future. Ask what they learned and what they might change next time. Encourage students to try it again later with different topics and see what new stems they pick up along the way.

Intermediate - Building Vocabulary and Practicing Polite Commands

Any time students interact with a chatbot, they are making language. Here’s an example of where you can have them double down by building some intermediate vocabulary, but also review the command form. I always do my best to practice speaking politely to the bots — not so much because I’m scared they’ll take over the world and want them to take pity on me (that’s only about 7% of it), but mostly because it’s always good to practice being polite and it tends to make me think more carefully about my language choices — not a requirement to work with chatbots, but a benefit to improving my own English.

Procedure

Research and Vocabulary Collection

Start by reading a short article or story in English. This could be online, in a textbook, or from any reliable source.

While reading, students should identify and list 10 words that they find challenging or unfamiliar.

Optional: Use the Chatbot to find meanings, synonyms, antonyms, and example sentences for each word. Alternatively, you may ask your students not to find the meanings yet as their prompts may involve discovering the meaning of the words through exposure.

Constructing Prompts

With the 10 words you've collected, the student’s task is to create five unique chatbot prompts. Ensure each prompt incorporates at least two of the challenging words.

Remember, the goal is not just to use the word but to construct meaningful, interesting, or engaging prompts that would lead to comprehensive replies from the chatbot.

Beginner Example: If the words are "counterfeit" and "envious," a prompt might be: "Can you please write a dialogue between two people and include the words ‘counterfeit’ and ‘envious’?"

Students will probably start with something like this, which is fine, but may not go into the depth that you’re looking for when you want students to explore their vocabulary. It may benefit them to get a couple of examples of language that can push their boundaries:

Story Starters: Students can craft the beginning sentence or two for a story that incorporates the challenging vocabulary. Their goal is to intrigue the chatbot to continue the story using those words without explicitly telling it that they are hoping it will repeat the key words

Intermediate Example: I am going to write the first few sentences of a story, and I’d like you to complete the story. "I stared across the bedroom at my husband of 16 years, then looked down at my beautiful and treasured engagement ring. A counterfeit. I was envious of the life I should have had."

Note that the goal is to eventually get students to come up with their own ideas for prompts that show a deeper understanding of the vocabulary, and not just the ideas you provide. Let them loose and see what they come up with!

Analysis

After the chatbot creates its responses, students should work their way through to see how successful it was. In the beginner example, did it include the words they wanted? Did the dialogue make sense? Was it interesting? In the intermediate example, students should read through the output, looking for repetitions of the key words. You  might spice this up by seeing who got their chatbot to reproduce the words the most. You also might ask them to search for examples of synonyms or terms that imply an understanding of the key words they were looking for.

During all of this, of course, you can double check students’ use of polite commands:

    • Can you please…
    • I’d like you to…
    • Would you mind…
Practice

After students are finished, you may give them a homework assignment to do a dramatic reading of their favorite story. They could record on Padlet or Flip for their classmates to listen to, which could in turn be used for even more future activities.

Advanced - Building Games With “If/Then” Conditionals

Many of us spent countless hours of our childhood reading through books where we had to make decisions like whether to send the main character down into a creepy basement or outside into the mysteries of the night. Building a “Choose Your Own Adventure” with a chatbot is the perfect realization of the opportunity to create this kind of game and to use conditionals to create boundaries in the game.

Procedure

Introduction

Begin by discussing the concept of game parameters and rules. Explain that games function based on a series of conditions and outcomes.

Show examples of simple game rules expressed as "if/then" statements, e.g., "If a pawn moves forward in chess, then it cannot move backward to return to its previous position."

Understanding Advanced Prompts

Showcase a few advanced prompts that have been used with chatbots to set up games or interactive stories. Here’s an example of a medium-length prompt I built (which you are welcome to test with your students). Note that it’s more advanced than what you’re trying to build here, but it can give a sense of where things might go.

Discuss the importance of clarity and specificity in these prompts.

Brainstorming Game Ideas

Ask students to think of simple game ideas or stories that can be played out with a chatbot.

Encourage creativity but remind them the core mechanics should be expressible via "if/then" statements.

Formulating Advanced Prompts

In pairs or small groups, learners should begin constructing the "if/then" statements that will form the backbone of their game's mechanics.

For a game idea like a treasure hunt, a statement could be: "If the user chooses the left path, then they encounter a river."

Have students combine their "if/then" statements into a cohesive advanced prompt to set up the game for the bot. For example:
"You are guiding a user through a treasure hunt. If they choose the left path, then they encounter a river. If they choose the right path, then they find a cave. If they ask for a hint, then provide a riddle."

Testing the Game With the Chatbot

Students input their advanced prompt into the chatbot, setting up the game parameters.

The next step is for the students to play out their game, interacting with the chatbot based on the rules and scenarios they've set up.

Encourage peer testing, where one group tests the game of another group, to get diverse feedback.

Review and Iteration

After testing, groups should discuss the outcomes of their games. Were there any unexpected responses from the chatbot? Were the rules clear?

Based on feedback, students should iterate on their advanced prompts to refine their games.

Sharing and Reflecting

Let pairs or groups showcase their games to the class. They can share their prompts and have each other play and then get collective feedback on how to make the games even better.

Before the next class, encourage students to play each other’s games and find ways to improve on them. Have them come to class ready with some statements like “If Mikhail and Firooza’s treasure hunt game included options to get across the river, then they could connect it to our Haunted House game.”

Conclusion

Once we start looking at talking to chatbots as the activity, then working to get it to do our bidding, we can see that there is a huge opportunity for students to stop thinking about language learning as “language learning” and instead to think of it as a challenge to their critical thinking and to learning how to build.

Your students are no longer relegated to passive learning as “students” — they are now Prompt Engineers in their own language acquisition journey!


Have you used prompt engineering as a way to help your students build their language skills? Please share what you’ve done in the comments below — we’re all better together, and there’s so much to learn and share here!

About the author

Brent Warner

Brent Warner is a professor of ESL at Irvine Valley College in California, and an educational technology enthusiast. He is co-host of the DIESOL podcast, the only podcast with a specific focus on EdTech in ESL. He frequently presents on the crossroads of technology and language learning, focusing on student engagement and developing learner autonomy. Brent likes his coffee black and his oranges orange. He can be found on LinkedIn at @BrentGWarner.

comments powered by Disqus

This website uses cookies. A cookie is a small piece of code that gives your computer a unique identity, but it does not contain any information that allows us to identify you personally. For more information on how TESOL International Association uses cookies, please read our privacy policy. Most browsers automatically accept cookies, but if you prefer, you can opt out by changing your browser settings.