Teaching languages in the age of AI: what if the real challenge is preserving cognitive effort?

I still vividly remember two language teachers: one in English, who had us listen to songs and discuss their themes, and another in Portuguese, who made us react to current events and debate our views on the Lusophone world. They both knew how to capture our attention by anchoring it in emotion and our connection to the world.

Today, language departments face a new challenge: how can they maintain this quality of cognitive interaction when students rely on artificial intelligence thinking they are learning when, in fact, they are not truly engaged?
AI translates, reformulates, corrects and does so increasingly well. But it does not create that cognitive tension, that moment of doubt, searching, and adjustment.

What language departments have been doing long before AI

Even before the term neuroeducation became popular, language teaching was already based on several of its key principles.

Attention

Learning to listen to a foreign language means learning to filter sounds, identify accents, and detect relevant cues in a stream of speech. This is what Stanislas Dehaene describes as the active selection of useful information. Language teachers stimulate this auditory discrimination through guided listening and comparison.
Artificial intelligence, when it simplifies or automatically transcribes, risks short-circuiting this essential cognitive activity.

Feedback

Every spoken attempt is a hypothesis the brain tests, adjusts, and reformulates a monitoring process (Vandergrift & Goh). Language teachers facilitate this process by redirecting, rephrasing, and providing immediate feedback.
AI can produce highly accurate corrections, but it remains an external agent, it does not generate metacognition. That remains the teacher’s role: helping learners become aware of their mistakes, understand their origins, and situate them within a specific cultural and learning context.

Emotion and memory

A word is never neutral: it’s tied to a situation, a culture, an identity. This emotional memory is essential for durable knowledge consolidation (Immordino-Yang & Damasio).
AI can correct syntax, but it cannot provoke surprise or connection those emotional sparks that deeply anchor linguistic learning.

AI as a reminder of the human value in language teaching

Artificial intelligence does not threaten language education it highlights its human value. It cannot recreate doubt, meaning negotiation, or the shared emotion that fuels linguistic exchange.
Teachers know that meaning is built in interaction, not perfection. It is within this shared imperfection this space of cognitive uncertainty that learning becomes alive and lasting.

At the departmental level, the challenge is no longer just mastering technology but preserving this relational and cognitive intelligence within increasingly automated systems. Three levers can support this goal:

1. Foster shared cognition practices

Interactive tools can make collective reasoning visible and ground discussions in learning processes rather than outcomes.
Teams can use them to jointly analyze participation dynamics, comprehension, or learner progress.

2. Move from a culture of correction to a culture of reflection

In the age of AI, the goal is no longer to give the right answer but to question the answer.
Encouraging comparison and peer discussion strengthens metacognition and helps harmonize practices without uniformity.

3. Preserve the emotional and social dimension of learning

Attention is also a collective phenomenon. Keeping moments of interaction, even brief ones, restores cognitive presenceand engagement. Technology thus becomes a support for synchronization, not a substitute for human connection.

Preserving relational intelligence in an automated world

Language teachers know that learning is not just the transmission of rules it emerges from the relationship between a word, a culture, a gesture, and an emotion.
Today, language program coordinators stand at a turning point. They have the opportunity to preserve what makes their pedagogy unique: attention, metacognition, and emotion, while integrating technological tools that make it possible to do so across diverse groups (where CEFR levels vary within the same class) and complex contexts (such as hybrid learning).