As much media coverage reveals, generative artificial intelligence (AI) technologies present many possibilities in education, alongside possible harms. From plagiarism use cases to the perpetuation of misinformation through hallucinations, or the exploitation of young women through deepfakes, the ethical and moral questions surrounding AI abound.
And yet, many educators are finding that these technologies, when used wisely and intentionally for good, hold the power to enhance accessibility and equity in education—for example, by facilitating multimodal content generation that better enables Universal Design for Learning (UDL) and supports the core tasks of English Language Learning as well.
In the fall of 2021, 5.3 million students in the United States alone identified as English Learners (ELs)—in some states, comprising over 20% of all public school students. And, according to data published by the Department of Education and Office of English Language Acquisition, less than half of public school teachers report taking one course on EL instruction during their post-secondary training, and only 5% hold a major, minor, or certificate in EL instruction.
There is, then, a pressing need for tools that better serve and support this population of learners in the U.S. and around the world, and that scaffold teachers’ instructional differentiation in the absence of formal training. Many practitioners are writing about how they’re effectively using generative AI to achieve these goals, highlighting a breadth of tools and use cases for them; we’ve compiled a selection in this round-up article to amplify the voices of thought leaders in this space and to jumpstart your practice in turn:
“Using AI Image Generators in World Language Classes”
By: Rachel Paperone
Published with: Edutopia
In this article, Paperone showcases teaching strategies from her World Language classes that are transferable to EL education—particularly, the use of AI image generation tools to complement the introduction of new vocabulary, providing multimodal representation and engagement (two tenets of UDL) to students. Paperone writes:
“I recently experimented with… asking my learners to use… image generators to create visual depictions of a new set of vocabulary. Given world language teachers’ emphasis on multiple literacies, I chose to implement this lesson with a notoriously tricky set of vocab for beginning language learners: daily routine. What seems like a simple, even boring, vocab set is difficult for students learning Romance languages because it introduces reflexive pronouns… I provided my learners with three basic pieces of information: a theme for the new vocabulary (daily routine); a list of 12–14 basic verbs in the first-person singular; and a goal: using Adobe Firefly to create a set of images that demonstrate the meaning of each expression.”
“Using AI to Support Multilingual Students”
By: Tan Huynh
Published with: Edutopia
In this piece, Huynh outlines utilizing AI in an EL context to generate mentor texts; align those texts and other course content with designated, differentiated Lexile levels; and to provide students with definitions for, and descriptions of, novel vocabulary.
Regarding each use case, Huynh writes:
- Definitions and descriptions: “When my students needed to know about the difference between a dictatorship and a democracy, we used Perplexity to produce a simple definition... We went further by prompting Perplexity to provide a list of advantages and disadvantages of each form.”
- Mentor texts: “I often use MagicSchool to generate a mentor text. I simply identify the content and the type of text (e.g., poem, news article, short story, review, op-ed, biography). In seconds, MagicSchool produces a satisfactory mentor text aligned with the criteria I identified.”
- Lexile Levels: “Eighth graders were expected to read a magazine article about the Willow project that was written at the 11th-grade level. I simply took the URL link for the article and gave it to Poe. I set the reading level to grade 5… Poe produced a much more accessible article...”
“AI Tools for the Four Skills in English Language Learning”
By: Matthew Sussmann
Published with: TESOL Connections
Sussmann highlights a wealth of AI tools for listening, speaking, writing, and reading tasks — from transcription and translation services to programs that provide feedback on students’ spoken content.
Specifically, Sussmann cites the following AIs as most helpful in the above: Speechify, Otter.ai, Readlang, FLOW Speak, Grammarly, ProWritingAid, Slick Write, QuillBot, and AI email writers like MailMeteor.
Allowing students’ scaffolded use of these tools in general education courses is one way to make instruction more accessible. Additionally, in the context of the EL classroom, integrating these tools can provide teachers with more time for individualized support, such as through station rotation activities in which AI programs provide feedback on students’ conversations while the instructor facilitates pull-out check-ins with students throughout the class period.
“How to Use AI Tools to Support English Language Learners”
By: Larry Ferlazzo
Published with: Education Week
Here, Ferlazzo again emphasizes the utility of AI’s ability to enhance the multimodality of instructional content, better serving ELs and aligning with UDL principles to make learning more accessible and retainable.
Ferlazzo writes:
“Text modification and the creation of multimodal resources are two of the areas where GenAI can provide invaluable support for language teachers. For instance, with tools like Elevenlabs for generating synthetic voices, and a range of image-, animation-, and video-generation tools such as Dalle-3, Adobe Express, D-ID or ArtFlow, we can create diverse multimodal materials and activities. Achieving this without such technology would be extremely challenging, if not impossible.”
Across articles, these accounts demonstrate how we can use AI as a teaching assistant, of sorts, to diversify the types of content we provide to students; the levels at which we provide that content, disrupting monolithic instruction and centering differentiation; and freeing up time and space during a class period to provide more one-on-one, individualized instruction. These practice are helpful in any classroom, but particularly aligned with equity-centered, accessible EL education.
Have other ideas or experiences using AI to support English Learners in your educational context? We’d love to hear from you at educators@writetheworld.org.
