Language professors grapple with AI's place in the classroom
Gabriel Guillén, a professor at the Middlebury Institute of International Studies, told a room of foreign language instructors at Washington University on April 2 that AI should supplement language learning, not replace it.
The talk, part of the Ginger Marcus Foreign Language Learning Speaker series, centered on a practical question: how should professors teach when their students have access to tools that can translate, generate text, and simulate conversations?
Guillén opened by polling the audience on their comfort with AI. Most fell between "cautious" and "curious," with two professors identifying as "committed."
Why humans still learn languages
Guillén argued that language teaching is fundamentally human work. Humans create unlimited meaning from limited sounds-a capacity no other species shares. But that capacity only persists when one generation passes language to the next.
"Teaching language is what makes us human," he said.
Despite the existence of translation tools like Google Translate, enrollment in language classes has grown. Spanish enrollment in American classrooms has doubled since the 1970s, from 250,000 students to 500,000 today. Migration driven by climate and economic pressures accounts for some of this increase.
Translation software cannot replace actual fluency. Learning a language creates identity, belonging, and trust-things that matter when moving to a new place or welcoming newcomers.
AI is different from past tools
Guillén asked the audience to match historical tools with what they expanded: the telescope extended human sight, the hammer extended hands, books extended memory. What does AI extend?
One professor said AI is fundamentally different because it produces something without clear authorship. The output comes from collective knowledge programmed by others, not from the user's own thinking.
Another noted that AI can replace certain human roles-it functions as a personal assistant in ways previous tools could not.
Redesigning assignments and policies
Rather than banning AI, Guillén suggested professors focus assignments on local events, audio or video work, and process over final product. This approach teaches students both how to use AI effectively and how to spot its mistakes.
He recommended building AI policies around four principles: agency, productive struggle, transparency, and voice integrity. These should evolve as tools change, not function as static rules.
"AI policy should be a process, not just something like a contract," he said.
One thing AI cannot replicate is immersion. Chatbots can simulate conversation practice, but they cannot replace study abroad experiences where students navigate real social and cultural contexts.
Skepticism remains in the field
Paolo Scartoni, a lecturer in Italian at Washington University, attended the talk hoping to learn how colleagues were adapting. He remains unconvinced that AI helps students.
"I'm a big fan of taking things slow," Scartoni said. "I think there's value in enjoying the journey. And I don't think there's always value in finding shortcuts."
He worries that AI removes a crucial part of learning: making mistakes and receiving feedback. His own experience with AI illustrates the risk. Since using AI to check his English emails, he has stopped paying attention to feedback from colleagues-he simply relies on the tool.
Scartoni also raised concerns about who controls these systems. "Behind the algorithm there's people, and that's scary because we are delegating our thinking, our learning, our writing, to some faceless people who are not in learning institutions. They are in private companies. Their goal is to make money."
Professors have a responsibility to protect students' thinking and to consider what using AI might do to their development as learners, he said.
The Romance Languages and Literatures department uses AI inconsistently across courses. This variation, Scartoni said, allows students to experience different teaching approaches-though it also means no consensus has emerged on best practices.
For educators working with AI for Education and AI Translation Courses, the challenge is clear: integrate new tools without sacrificing the slow, deliberate work that builds genuine language competence.
Your membership also unlocks:
