I have shared many columns about how teachers might want to consider—and how they might not want to consider—using artificial intelligence in the classroom.

You can see many of those previous posts here. You can also see many more ed.-related resources here, including two animated videos about AI in education that Katie Hull and I did with Education Week.

Today’s post focuses on using artificial intelligence with English-language learners (as did a previous one).

Few teachers have had more experience applying AI tools with ELLs than Svetlana Kandybovich, and she agreed to contribute to today’s post. Svetlana includes links to free AI tools she has created for teachers to use with students.

Opportunity or Remix?

Svetlana Kandybovich is a teacher, teacher trainer, and materials writer with more than 20 years of international ELT (English-Language Teaching ) experience. She shares her insights on teaching, learning, and professional development with fellow educators and teacher trainers through her blog, ELTcation.

Do GenAI-powered tools offer new opportunities for language teaching and learning, or are they simply old school concepts remixed with AI technology?

The ease of creating GenAI-powered tools has led to an “AI rush” among edupreneurs, offering various tools promising to revolutionize language learning by making it more engaging, personalized, adaptive, and inclusive. However, as with any other new or emerging technology, without proper integration into our practices to ensure it truly adds value to learning, this potentially powerful technology might just end up being wasted.

What Makes GenAI Effective?

A tool is effective when it adds value to learning. Following the SAMR approach, the impact of a new tool is higher when it transforms learning by redesigning or redefining the pedagogy and introduces new possibilities that were previously unimaginable. Put simply, a tool is most valuable when it is indispensable and helps teachers create a better learning environment than any other methods or tools available.

When used effectively, every tool has the potential to improve learning and learning outcomes, but it’s the teacher’s consistent and principled approach that truly makes it work.

Image by Svetlana Kandybovich

Key Elements

When it comes to effectively using and integrating GenAI, three things are key: the teacher’s expertise, understanding GenAI, and being creative. They are all tied to how GenAI works, which can be likened to a game of chance. While we can manage the input, no one has complete control over what comes out of GenAI (and likewise, nobody has fully understood its capabilities yet). This highlights the importance of teachers being experts in their field.

To achieve quality results, teachers should have a clear idea of their desired output and the ability to evaluate, iterate, and fine-tune it further. Teachers’ AI literacy is essential for addressing the limitations and risks associated with GenAI, while creativity is needed to push boundaries, explore new activities and tasks, and unlock the technology’s potential to make learning even better.

GenAI Use Cases

We can discern quite distinct ways in which language teachers use GenAI by studying their experiences shared in blog posts, social media groups, forums, and during webinars and conferences.

The largest group by far involves replicating and automating existing practices. Many examples of this can be found, ranging from direct use of large language models (LLMs) or single solutions and apps like ChatGPT, Copilot, or Gemini, to various platforms built on top of LLMs designed for educators of various subjects, such as MagicSchool or RockettAI, and for language teachers in particular, such as Twee or LingoTeach. These platforms enable teachers to create a wide range of worksheets, develop lesson plans, write texts on any topic and for any proficiency level, and design various types of exercises. They leverage the capabilities of LLMs, particularly their superfast text generation. The idea is to save time, but speed isn’t everything.

Sometimes, just like with any other third-party tool, your pedagogical approach may differ from that of platform developers. Therefore, what these platforms churn out might not fit your needs perfectly. Without the right customization options, they might only provide a quick solution that doesn’t truly benefit your learners, leaving you needing to invest more time to fix things up.

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.

The impact level increases when tools are customized to meet specific needs and provide teachers with customization functionality, allowing them to incorporate their unique contexts and experiences. For instance, I have developed two tools to aid language teachers in brainstorming and designing personalized activities for their learners.

The Five-Minute Activity Generator helps language teachers come up with quick activities for the classroom, while the Role-Play Generator is crafted to generate role-play scenarios and role-play cards tailored to learners’ needs and learning objectives. Both tools tap into the content-generation powers of LLMs and are great for sparking ideas. But what really makes the magic happen is when teachers put their own twist on things, adding their personal touch and context details.

One novel use case involves the creation of AI-powered chatbots using various GenAI tools to simulate real-life interactions and offer new learning experiences. These chatbots can be designed as both stand-alone activities and as part of specific lessons. You can assign a particular persona to a chatbot, whether it’s an imaginary character or a famous person, and it will act accordingly. For example, in the Tumbleweed Invasion lesson, learners engage in dialogue with Alex the chatbot to uncover the story’s details as narrated by Alex.

I have created and tested several chatbots to support my learners, each with different objectives that define their behavior and language. For instance, I’ve developed ThinkChat Buddy to assist my intermediate and upper-intermediate learners in practicing conversation skills across various scenarios. Essentially, it acts as a conversation scaffold, guiding learners through dialogues by thinking out loud, explaining why it responds as it does, clarifying expectations, and providing gentle and encouraging feedback. This chatbot is not a replacement for classroom conversation practice; it’s a supportive tool that extends learners’ practice, preparing them for real-life conversations.

Two important elements of using chatbots effectively in language learning are their design and monitoring. To prevent chatbots from becoming just one-time engagement tools, it’s essential to align their behavior, language, and instructions with learners’ specific needs, proficiency levels, and learning objectives. From my observations, even proficient learners often struggle with unfocused chatbot interactions, when they are simply instructed to “go and chat,” resulting in minimal learning value. Additionally, circling back to the issue of control over GenAI output and its associated risks, it’s essential for teachers to closely monitor learner-chatbot interactions and quickly adjust chatbot tasks and behaviors as needed.

As GenAI continues to advance, we will discover even more new ways it can enhance and transform language learning. Yet, its effectiveness truly depends on the skill level of the teachers, their understanding of AI, and their ability to think creatively.

Consider contributing a question to be answered in a future post. You can send one to me at lferlazzo@epe.org. When you send it in, let me know if I can use your real name if it’s selected or if you’d prefer remaining anonymous and have a pseudonym in mind.

You can also contact me on Twitter at @Larryferlazzo.

Just a reminder; you can subscribe and receive updates from this blog via email. And if you missed any of the highlights from the first 12 years of this blog, you can see a categorized list here.