ESL/ELL Students and AI: Teaching Language Learners in Multilingual Classrooms

The Multilingual Classroom: Reality vs. Preparation

In the 2025–2026 school year, approximately 10.4% of U.S. public school students are classified as English Language Learners — over 5 million students speaking more than 400 home languages. Yet the majority of general education teachers have received minimal preparation for ELL instruction. Research by Gándara and Maxwell-Jolly at UC Davis found that less than 30% of teachers with ELL students in their classrooms reported feeling adequately prepared to serve them. The gap between who is in the classroom and what teachers know about serving them is one of the most consequential preparation failures in American teacher education.

This guide does not assume you have an ESOL endorsement. It assumes you are a general education teacher with ELL students in your class, you want to serve them effectively, and you need practical, research-grounded strategies that you can implement within the constraints of a full teaching load serving 25–30 students with a wide range of needs.

The WIDA Framework: Speaking the Language of ELL Proficiency

The WIDA English Language Proficiency Standards provide the most widely used framework for understanding and communicating about ELL student needs in U.S. schools. WIDA defines six proficiency levels:

• Level 1 — Entering: Very limited English; relies heavily on visual support, gestures, home language
• Level 2 — Emerging: Can communicate basic needs and understand simple sentences with support
• Level 3 — Developing: Can understand and produce simple and some complex sentences; errors common
• Level 4 — Expanding: Can communicate in most contexts with some difficulty in academic language
• Level 5 — Bridging: Near-native social language; academic language developing across disciplines
• Level 6 — Reaching: Comparable to native speaker performance; may still have discipline-specific gaps

Understanding your students' WIDA levels (typically reported on ACCESS for ELLs test results) allows you to calibrate the language demands of your instruction appropriately. A Level 1 student and a Level 4 student may both be on your ELL roster, but their needs are categorically different — treating them with identical accommodations serves neither well.

Translanguaging: The Research-Backed Asset Approach

Traditional approaches to ELL instruction have often been deficit-oriented: students speak a language other than English, which must be managed or overcome. Research over the past two decades has fundamentally challenged this framing. Ofelia García's translanguaging research at the City University of New York documents that multilingual students' home languages are cognitive assets — not competing systems to be suppressed, but resources that support comprehension, thinking, and English language development.

Translanguaging pedagogies allow students to use their full linguistic repertoire during learning — reading a text in English while discussing it in the home language with a bilingual peer, writing notes in the home language before composing in English, using bilingual dictionaries as bridges. Research by Celic and Seltzer (2011) and García and Wei (2014) shows that classrooms that embrace translanguaging produce better content learning, stronger academic language development, and more positive language identity outcomes than classrooms that enforce English-only policies.

Practical implementation: create structured opportunities for bilingual peer talk (pair ELL students who share a home language for initial comprehension checks), allow home language note-taking during instruction, and use bilingual word walls that include home language translations alongside English academic vocabulary.

The BICS/CALP Distinction: Why Socially Fluent Students Still Struggle

Jim Cummins' distinction between Basic Interpersonal Communication Skills (BICS) and Cognitive Academic Language Proficiency (CALP) is perhaps the most important concept for general education teachers to understand about their ELL students. BICS — the language of casual conversation, playground interaction, and social navigation — typically develops within 1–2 years of consistent English exposure. Students who have been in English-speaking schools for 2–3 years often appear fully English-proficient in social contexts.

CALP — the language of academic texts, disciplinary discourse, and complex written argumentation — typically requires 5–7 years to develop to grade-level proficiency. An ELL student who has been in U.S. schools for three years may be socially fluent but still significantly below grade level in academic English. This distinction explains a pattern that frustrates many teachers: the student who seems to understand everything in casual conversation but can't independently read a grade-level text or write an academic essay. They're not being lazy or not trying — they're still developing a language competency that takes years longer than most teachers expect.

Sheltered Instruction: The SIOP Model

The Sheltered Instruction Observation Protocol (SIOP), developed by Jana Echevarria, MaryEllen Vogt, and Deborah Short, provides a research-validated framework for making grade-level content instruction accessible to ELL students without watering down the content. SIOP's key components include:

• Explicit content and language objectives: Every lesson should have both a content objective ("Students will explain the causes of the American Revolution") and a language objective ("Students will use cause-and-effect language to explain: 'because,' 'as a result,' 'led to'")
• Building background knowledge: ELL students may lack the U.S.-specific cultural background knowledge that academic texts assume. Pre-teaching relevant background and connecting to students' own cultural knowledge is essential
• Comprehensible input: Speech rate, vocabulary complexity, and sentence structure should be adjusted to student proficiency level; visual support, gestures, and demonstrations should supplement verbal instruction
• Student interaction: Structured opportunities for ELL students to practice academic language — not just listen to it — are essential for language development
• Review and assessment: Regularly checking comprehension through non-verbal and verbal means throughout instruction

AI Tools for ELL Instruction: Opportunities and Cautions

Real-Time Captioning

AI-powered real-time captioning (available in Google Slides, Microsoft Teams, and Zoom) makes spoken instruction accessible to students who process written English more readily than spoken English — common at lower WIDA proficiency levels. Captions also serve as a visual record of instruction that students can reference during activities. The quality of AI captioning has improved dramatically — current error rates for clear speech are below 5% for major languages.

AI Translation for Comprehension Scaffolding

AI translation tools have become genuinely functional for most major languages. Using translated texts as initial comprehension scaffolds — reading the home language version to establish meaning before engaging with the English version — reduces the cognitive load of simultaneously learning content and language. The caution: translation should scaffold toward English engagement, not replace it. Students who always read translated content are not developing the English academic language they need.

AI Writing Assistance for ELL Students

AI writing assistants can help ELL students move from ideas to coherent English prose — reducing the gap between their conceptual knowledge and their ability to express it in academic English. The pedagogical key is positioning AI assistance as a drafting scaffold that the student then revises with growing independence, rather than as a finished product generator. Grammar error feedback (without automatic correction) supports language development; automatic correction without explanation does not.

Cultural Responsiveness Caution

Many AI tools trained primarily on English-language, U.S.-centric content may produce responses that are culturally inappropriate or irrelevant for students from different cultural backgrounds. Review AI-generated content for cultural assumptions before using it with ELL students. AI cannot currently provide the culturally responsive teaching that requires human knowledge of students' backgrounds, communities, and cultural assets.

Family Engagement When Language Is a Barrier

ELL students' families are often their most powerful academic assets — and schools' most underutilized partners. Research by Henderson and Mapp consistently shows that family engagement is one of the strongest predictors of ELL student academic outcomes, independent of language proficiency. The barriers to engagement — language distance, cultural differences in the expected role of family in schools, work schedules that prevent attendance at school-day events — are real but reducible.

AI translation tools have made multilingual family communication more accessible than ever: written communication translated into home languages (with professional translation for high-stakes documents), multilingual school websites, and AI-assisted interpreter services for phone communication. Many schools have found that community liaisons — staff members who share cultural and linguistic background with the largest ELL populations — are among the highest-ROI investments for family engagement.

Ready to see the difference? Try Koydo free today →