The Future of Technology in English Learner Instruction, Part 1

 

Now, let’s get into the specific usage of AI to support the work you're already doing with your English learners.

Adapting Content for Different Proficiency Levels

Making grade-level content accessible to English learners of different proficiency levels is essential. This should also be the responsibility of both ELD and content teachers. Unfortunately, providing appropriate scaffolding and differentiation for multiple levels is extremely time-consuming. Moreover, many content teachers don't know how to do this work effectively.

This is where AI becomes valuable for both ELD specialists and content teachers. Here’s an example: A 9^th-grade science teacher has students studying ecosystems. Their proficiency level 4 English learners can handle the text in their textbooks, but their proficiency level 1 students need significant support. To help these students access the text on ecosystems, the science teacher uses this prompt in ChatGPT:

"Rewrite this passage about ecosystems at an [appropriate Lexile level] for an English learner. Use simple sentences, present tense verbs, and high-frequency vocabulary. Avoid idiomatic expressions and figurative language. The passage is: [paste original text].”

Within seconds, the LLM responds with a properly scaffolded version of the text for the proficiency level 1 student. The answer may be exactly what you are looking for, or it may be 80% there but requires some refinement.

The key to getting useful results is to be specific in your prompt and to iterate. For example, tell the LLM exactly what proficiency level you need, what language features to include or avoid, and what content must be preserved. If the answer is not what you expect or falls short, ask follow-up questions or prompts.  Feel free to disagree with the answer with your reasons why and ask for a better response.

If you're coaching content teachers on supporting their English learners, teaching them to use AI for differentiation gives them a practical tool they can implement right away.  AI is not a substitute for them trying to understand their students' language needs, but it removes the "I don't have time" barrier.

Supporting Comprehensible Text

Your newcomers are building English skills, but they still need to understand the content. AI can bridge that gap by supporting comprehensible text.

ChatGPT can translate text into the L1 language, but it can go further than simple translation by incorporating comprehensible text strategies.

Try this prompt:

"Explain the concept of photosynthesis in Spanish at a middle school level. Then provide the same explanation in English with Spanish cognates bolded."

This gives your Spanish-speaking students access to content while helping them build connections to English academic vocabulary. You can also ask AI to summarize challenging texts in their home language, partially translate (translating only challenging words or phrases), or create visual images that help students understand concepts.

Microsoft Edge's text-to-speech feature adds another layer of support. Students who aren't yet literate in their home language can hear translated text read aloud. Navigate to any webpage, PDF, or document in the browser, right-click, and select "Read aloud." You can adjust the speed and choose from different language voices. This turns written translations into accessible audio support.

Creating Instructional Materials and Activities

AI can analyze texts and generate teaching materials faster than you can create them manually. Here's where the real time savings happen.

Start by asking ChatGPT or another LLM to analyze a text for vocabulary and phonics patterns:

"Analyze this passage for Tier 2 vocabulary words that would be challenging for English learners. Then identify any words that follow the CVCe pattern."

The LLM can pull out academic vocabulary and phonics teaching opportunities you might have missed when skimming the text.

Once you have your target words, ask the LLM to create activities based on these words:

"Create a vocabulary matching activity using these 10 words. Include the word, a student-friendly definition, and an example sentence for each."

You can use these types of prompts with any LLM.  However, if you want to try an AI tool specifically designed for teachers, you might want to check out Diffit (www.diffit.me). Upload any text or paste a URL, select your students' reading level and language, and Diffit automatically creates a simplified version along with vocabulary lists, comprehension questions, and even discussion prompts. It's particularly useful when you need to quickly adapt current events articles or authentic texts to a specific reading level.

Generating Questions, Scaffolds, and Assessments

Checking for understanding throughout class time is a given, but creating effective questions and scaffolds requires careful planning. AI can help generate both to match your instructional goals.

For class discussions, start by creating questions that promote critical thinking:

"Create 5 open-ended discussion questions about this text that require students to make inferences and support their answers with evidence."

English learners often know what they want to say, but struggle with the academic language needed to express their thoughts and logic. Providing some initial structure can help boost confidence.

So, follow up with a prompt that asks for sentence starters to pair with those discussion questions.

"Create 6 sentence starters that students can use to agree, disagree, or add to a classmate's idea during a discussion about climate change."

The same principle applies to writing tasks. Use AI to generate the sentence frames students need before they begin:

"Create sentence frames for a 5-paragraph essay comparing two characters. Include frames for the introduction, body paragraphs, and conclusion."

Another use case for AI is to create ad-hoc formative assessments. AI can create multiple question types quickly based on a sample text:

"Create 8 multiple-choice questions about this passage. Include 4 questions about main ideas and 4 about supporting details. Provide an answer key."

Make sure you review the questions and answers for relevance and accuracy (always quality-check AI output). Adjust the suggested assessment as needed, and you're ready to go.

Providing Instant, Value-Added Feedback on Student Writing

Timely feedback matters, but when you're reading 30+ student responses, it's hard to give individualized comments with detailed suggestions for improvement. AI can help you provide more value-added feedback faster.

Copy a student's paragraph response into ChatGPT with this prompt:

"This is an opinion paragraph written by a proficiency level 2 English learner. Identify strengths in their writing and opportunities for improvement. Provide at least one specific revision to improve clarity. Point out grammatical errors. Do not rewrite the paragraph."

AI will give you feedback points you can share with the student. You still make the teaching decisions about which feedback to prioritize, but AI helps you see patterns and possibilities you might miss when you're tired at the end of a long day.

It’s important to note that we do not recommend using an AI tool or an LLM for summative assessment or grading. While it often gets close, AI is still not 100% accurate and should not be used to evaluate “high-stakes” activities.