Learning Reboot: Using AI Tools to Support Multilingual Learners

Dr. Jamie Lake, Oklahoma State University, Stillwater, Oklahoma, USA
Jessica Marine, Oklahoma City Public Schools, Oklahoma City, Oklahoma, USA

The development of Artificial Intelligence (AI) advanced enough to write coherent responses is becoming widely available, often with no cost, making access easier for users across the globe. This may initially seem concerning for educators. If AI can compose a paragraph or answer test questions, what is teaching and learning going to look like in a few years?

Only a couple of decades ago, teachers expressed similar fears about the world wide web; students could easily go there to find answers and even “borrow” papers. Yet many educators today would agree that teaching and learning are better because of the internet; we can edit collaboratively in cloud-based documents, create quizzes and games that score themselves, score student work digitally and enter grades without fear of spilling coffee on a pile of papers, and a million other things that are old hat by now. Once we as teachers learned how to use the internet, we began to harness the power of improved efficiency, collaboration, and productivity.

These days, we even have tools and frameworks for considering how we use technology in our classrooms, such as the SAMR model (Puentedura, 2009), which considers to what extent we are utilizing technology tools to enhance learning in the classroom. The idea of SAMR and similar models is that we should not be using technology to replace the same old routines, such as using an apple pencil and an iPad to take basic notes; instead, technology should elevate the experience, such as taking notes in a digital notebook, saving the notes in a folder organized by subject area, utilizing clickable links in a notes format, adding images from the internet to notes, and more. The internet, plus the thinking and frameworks developed by educators around it, have indeed elevated teaching and learning.

In many ways, AI will likely become the internet revolution of the modern era. We may be apprehensive about AI for similar reasons; it makes cheating easier, for example. But like the problems of the early internet, educators are already finding creative solutions. For example, test questions may be written as “AI resistant” where it’s more difficult for generative AI to answer correctly. In fact, for us as educators, we find these challenges to be advantageous, as they often challenge us to write questions that engage higher order thinking skills. As such, we see AI as an opportunity to both challenge ourselves as educators and reinvent classrooms as a more effective and supportive space for learners.

If you have used AI tools at all, you may have noticed the incredible ability generative AI has to use and produce language. With email alone, AI can compose a draft, rewrite to reflect a specific tone, summarize a thread, personalize an email for the recipient, suggest edits, and more. Teachers of multilingual learners may have understood something much bigger than email; after all, no one is writing my emails except me! The broader message here is that AI understands language very well! This makes AI a tool of incredible potential for educators supporting multilingual learners in the classroom.

Translanguaging

Even if you primarily use AI in English, AI itself is multilingual! This means AI tools can be excellent for bridging the gap, especially for students who are translanguaging or relying on more than one language to communicate. Translanguaging learners, especially newcomers, often have a wealth of knowledge that they may struggle to express in English. Tapping into a student’s native language can lower communication barriers and help students bridge languages faster.

Cognates

Cognates are one excellent way to support translanguaging learners. A cognate is a word that sounds the same and has the same meaning in two languages (bicycle/bicicleta in English and Spanish). Cognates can be magical because, once learners are aware of them, they have to do much less work to have the word in both languages. Take the bicycle/bicicleta cognates, for example. If you are reading this article in English, you likely know the word bicycle in English. If you didn’t know the Spanish word bicicleta, you probably aren’t comfortable yet using it in a sentence; however, you are likely to recognize it if it is spoken or written in front of you, which means it has already started to enter your comprehension vocabulary.

It is important to keep a few tips in mind for working with cognates.

• Language distance: Some languages will have more cognates with English than others. English is a Germanic language with Latin influence, so languages in these families are more likely to have overlap. A language like Spanish will have many cognates with English; a language such as Korean will have fewer.
   
• Prior knowledge: Cognates are only helpful for words students already know in their native language. For example, if we were in a classroom learning the word propincuidad in Spanish, it may not be helpful to be told that the English equivalent is propinquity (meaning closeness or proximity). This is because you must already know the word in your native language to make use of a cognate. Similarly, a student learning about photosynthesis for the first time might be unlikely to know that word in their native language, even if the term is a cognate.

AI is an amazing tool for utilizing cognates in the classroom. It can generate a list of words that are cognates, identify cognates in a passage, give background knowledge on cognates, and more. Try the following prompts with your favorite generative AI program:

• What are the Spanish cognates in this text? [paste in text]
• Which words in this sentence are Spanish cognates?
• Can you change the following sentence to include more Spanish cognates?

Language Transfer

While cognates are applicable to every classroom and subject area, understanding language transfer may be more specific to literacy and English language development courses. Understanding language transfer simply means comparing two languages, their similarities, and their differences. Understanding differences can be especially helpful in supporting multilingual learners in the early stages of language development and literacy.

For example, an English language arts teacher may want to understand more about nouns in Korean grammar before teaching a lesson on singular and plural nouns in English. By asking AI, he or she may be able to anticipate misunderstandings a native Korean speaker may have during the lesson.

Similarly, a teacher preparing for a lesson on consonant sounds in English may ask an AI program which consonant sounds in English are found in Spanish, and which ones are different or completely absent, in order to anticipate support Spanish speakers may need.

Tap into AI’s wealth of knowledge of comparing languages by trying the following prompts:

• What consonant sounds are the same between English and Vietnamese?
• What vowel sounds are the same between English and Vietnamese?
• How do noun rules compare between English and Vietnamese?
• How does the past tense compare between English and Vietnamese?

Understanding and Using Academic English

Most students who are translanguaging are newcomers, or new arrivals who are relying on their home language for communication at school. However, for most schools, the majority of MLs are non-newcomers, sometimes called long-term ELs. Umansky and Avelar (2023) upgrade the terminology to gifted multilinguals, explaining that, far from a monolith, students who have been under the EL label are a diverse group, often from vastly different backgrounds, with varying strengths and areas for growth in language and content that run the gamut like all students. However, under the EL label, this group has one primary commonality: they have yet to demonstrate English proficiency in a way that meets state exit criteria. For most of these learners, the beginnings of language, what Cummins (1979) dubbed as BICS, and commonly called social language, were mastered long ago. These learners are still working to master what many native speakers also lack - academic English (Umansky & Avelar, 2023), or what Cummins called Cognitive Academic Language Proficiency, or CALP. Academic language is a requirement for school success, not only in reading and composition, but in utilizing academic discourse verbally. As such, many learners across the school setting, including MLs who have mastered social language, need support in academic language. While there are many ways AI can assist in this support, we will focus on two.

Simplifying Text - Supporting Receptive Language

Students may need support in one or more of their receptive language skills, such as reading or listening. In the case of listening, educators can adjust or modify speech on the go. Reading is frequently more of a challenge for learners, partly because students may be used to listening all day, because reading requires skills such as decoding, and because there are often fewer nonlinguistic cues with reading (as opposed to gestures and visuals that may accompany listening to a speaker). Despite the challenge, it's important that students not avoid reading texts in the content areas. Instead of a textbook written on grade level, educators can utilize AI to create or modify texts that use academic language on a level more accessible for students.

We like sites such as Diffit.me or MagicSchool.AI for this purpose because we find they have targeted resources:

• Magicschool.ai
  • Text Rewriter Tool - Take any text and rewrite it in different words, change topics or characters - anything!
  • Text Leveler Tool - Take any text and adapt it for any grade level to fit a student's reading level / skills.
     
• Diffit
  • Adapt any reading, excerpt, or article for any reading level.
  • Generate short informational or narrative texts on any topic, for any reading level.

Identifying Academic Language - Supporting Expressive Language

Multilingual learners, especially at the upper levels of language development, are also likely to need extensive practice and use of academic language (AL) in expressive skills, such as speaking and writing, in all content areas. One of our favorite, most accessible tools for all classrooms is the sentence frame with interchangeable options. Sentence frames (plus sentence starters or talk moves) give students a model for using more advanced academic language, while still conveying the student’s own meaning. These may sound familiar, but here are some of our favorite AI resources for finding lesson-specific sentence frames:

• Try the following prompts with a general-purpose AI program:
  • What are some sentence frames for a lesson about…?
  • What are some sentence starters for a lesson about…?
  • What are some sentence stems for a lesson about…?
• Use the Sentence Starters Generator at MagicSchool.AI to generate grade level and topic-specific sentence frames or starters.

AI tools have an incredible ability to help teachers understand the native languages of multilingual learners for lessons, build in student supports, challenge all learners to excel with academic language, and more. There are undoubtedly hundreds of ways we will use AI in classrooms in the future which we have yet to even think of! AI is an exciting tool. Wariness and caution are always warranted; as with the early years of using the internet in classrooms, we should continue to be cautious with these tools, but we hope we can all continue to learn and explore how new tools can support teachers and learners in classrooms every day.

This article was not written by a robot.

References

Cummins, J. (1979). Cognitive/Academic Language Proficiency, Linguistic Interdependence, the Optimum Age Question and Some Other Matters. Working Papers on Bilingualism, No. 19, 121-129.

Puentedura, R.R. (2009). As We May Teach: Educational Technology, From Theory Into Practice. Apple.

Umansky, I. M., & Avelar, J. D. (2023). Canaried in the Coal Mine: What the Experiences and Outcomes of Students Considered Long-term English Learners Teach us About Pitfalls in English Learner Education… and What We Can Do About It. Journal of Education for Students Placed at Risk (JESPAR), 28(1), 122-147.

Jamie Lake is an assistant professor of K-12 English language learning at Oklahoma State University. She holds a Masters of Education in Bilingual/TESL from the University of Central Oklahoma and an EdD in Education Leadership from the University of Oklahoma. Jamie’s research interests include multilingual learners, academic language development, grade retention, and supporting experienced multilingual learners.
 

Jessica Marine works as an ELD program administrator specializing in secondary education with Oklahoma City Public Schools. She is also working on an educational leadership fellowship as an alum of Teach for America. Jessica holds a Masters of Education in Educational Administration from the University of Oklahoma.