I’ve been sharing the story of how interest in an innovative approach to language learning that we call Facilitated Interdependent Language Learning (FILL) was sparked four years ago and has been growing ever since. By chance, perhaps(?), FILL's journey has coincided with the explosion of interest in how Generative AI (GenAI), like ChatGPT and other Large Language Models (LLM’s), could transform language learning and teaching.1
As I described in my last Substack post, Taking FILL National, the FILL team members were early “explorers” of ChatGPT. We were very curious and immediately attracted to the potential benefits of tools like ChatGPT for personalizing the language experience of learners in a FILL setting. And for a FILL teacher/facilitator, who might be searching out resources for ten or more languages at one time, it could be invaluable. Even early on after ChatGPT was launched, we were surprised that English could be translated into less commonly taught languages like Romanian, Russian, and Ukrainian at all, much less with reasonable accuracy (at least for a first draft translation).2
In spring of 2025, I was invited to give a keynote at the FLEAT 8 Conference in Honolulu, Hawaii:
Language Learning in the Age of AI
As AI reshapes education, how can we harness its potential to create more equitable, personalized, and community-connected language learning experiences? This keynote introduces Facilitated Interdependent Language Learning (FILL)—a bold new approach where learners pursue diverse languages side-by-side, guided by language professionals in a facilitative role rather than a traditional instructive one. We’ll explore how FILL can extend access to underrepresented languages and empower learners to set goals, reflect, and collaborate while pursuing different languages. In the spirit of ‘ohana, participants will be invited to reflect on the new competencies language educators need—and the powerful, yet still to be determined, roles AI can play in building our global language learning family.
It was exciting to present about FILL at an international conference, focused on language learning technology. Here is a recording of my FLEAT 8 Conference Welcome & Opening Keynote: "Language Learning in the Age of AI"
The Age of AI: Possibilities are huge
My colleague Tom Welch is writing weekly on Substack3 now about learning, AI, FILL, and anything that will transform how we currently do “education.”
One lesson the Camino taught me: every journey is personal. There’s no single right path, no universal pace, and no ‘failure’ for not following someone else’s map. Our students—and our schools—deserve that same flexibility. (Quote and photo courtesy of Tom Welch)
The Age of AI: Friend and Foe?
I, too, am enthusiastic about the possibilities of GenAI. But I am also wary. In the FLEAT keynote, I shared a new book, AI Con: How to Fight Big Tech's Hype and Create the Future We Want.4 What drew me to this book was, first of all, the fact that one of the coauthors, Dr. Emily Bender, is a colleague from the University of Washington. Over the years, she has provided immeasurable support for our Japanese Dual Language Immersion program in Seattle Public Schools (which I helped found in 2001 and both her sons graduated from). So, I knew she had a strong commitment to language learning. But, as someone myself who worked in Information Technology for almost 25 years, I knew that the current AI bandwagon was not going to be “all good,” so I wanted to benefit from Emily’s and her coauthor, Alex Hanna’s, insight from being more directly involved. Dr. Bender directs the University of Washington Computational Linguistics Laboratory, and she specializes in computational linguistics and natural language processing.
There are so many insights from the book, but one thing that really struck me is the huge cost of building the kinds of data centers needed to support the rapid expansion of AI large language models (especially for tech billionaires chasing AGI–Artificial General Intelligence). This has an enormous impact on the environment, just at a time when our efforts to address climate change are under unprecedented attack. Of course, the funds to run these centers also reduce the potential funds to support other areas of great need in our country and around the world, such as education, hunger, housing, healthcare, and the list goes on.
I find myself occupying a place of ambiguity. (Tom Welch suggested that perhaps I needed to think of AI as Friend and Foe, rather than Friend or Foe.) I want to continue to be open to exploring the opportunities that GenAI presents for language learning, but I also want to always be cognizant that there are costs, both known and unknown. I’m hoping that human creativity (or maybe I should say, creative humans) will discover ways to mitigate or greatly reduce the negatives of using AI. One thing for sure, there is a “con” going on right now. The tech companies are not transparent at all about their products and how their large language models are constructed (using whose data?).
Next time: In the spirit of ‘Ohana, Fostering a Family of Language Learners. We’ll be exploring new implementations of FILL. Thanks for subscribing to this Substack. Please like and share it. I welcome your comments.
See, for example, this AI Resources page on the ACTFL website:
https://www.actfl.org/educator-resources/ai-resources (retrieved 8/16/2025), including “Making AI Work for Language Teachers.” https://www.actfl.org/professional-learning/discover-series/making-ai-work-for-language-teachers (retrieved 8/16/2025)
Although the translation project was not part of FILL, it was part of supporting an Open Educational Resource (OER) grant project for the Office of Superintendent of Public Instruction in Washington state. We were translating lesson plans developed by the World Affairs Council Global Classroom in Seattle, WA on the UN Sustainable Development Goals (SDGs) into languages that were part of a separate Heritage Language Grant, funded by OSPI.
UN Sustainable Development Goals - Global Competence Lessons
https://oercommons.org/courseware/lesson/105481/overview
The final translations (produced by human translators, but using first draft translations by ChatGPT) are found here: Lessons 1-6 in other languages:
Russian | Romanian | Spanish | Ukrainian.
Tom Welch in “Still Lighting Learning Fires”
https://substack.com/@twelchky
Age Inappropriate: Time To Light Some Fires (6/20/2025)
Bender, E. M., & Hanna, A. (2025). The AI con: How to fight big tech's hype and create the future we want. Harper. https://thecon.ai/
I appreciate that you remind us all that nothing comes without costs -- this is such a difficult area for me since on the one hand, I am amazed by, enthusiastic about and impressed by AI's ability to increase opportunities for language learning for so many more people -- students and adults. At the same time, while it has become ubiquitous, there are certainly significant ancillary costs for energy, etc that you have mentioned. And according to some sources I've read, many of these costs are borne by already marginalized communities. I hope you'll continue to raise this issue and I hope we can convene some heart-to-mind discussions about it. I think this is an issue that should be on the NCSSFL agenda as well as an agenda item for the Neural Network.