How language learners and Erasmus students use AI notes to get more from immersion

The fastest language learners are the ones who spend the most time inside the language — listening to it, speaking it, being corrected in it, surrounded by it. Every hour of immersion is worth more than any equivalent time with a textbook. This is why study abroad programmes, Erasmus exchanges, and language immersion experiences produce accelerating progress in ways that classroom instruction alone doesn't match.

The catch: immersion is cognitively exhausting, and the moments when you learn the most are also the moments when you can least afford to be writing things down.

A foreign-language lecture on a topic you're unfamiliar with, delivered by a native speaker at full speed, with idiomatic vocabulary you haven't encountered before, is one of the most linguistically rich environments you can be in. It's also one where taking notes in the target language while simultaneously processing spoken input is functionally impossible for most learners.

AI removes the constraint.

Why language immersion produces notes that don't help

A French student attending lectures at Sciences Po Paris, or a Spanish student at Carlos III in Madrid, faces a specific problem: the lecture is the learning environment. The vocabulary they encounter, the structures they hear, the academic register they're immersed in — all of this is the point.

Notes taken in the foreign language during an immersive lecture tend to be incomplete. The learner misses spoken words while writing previous ones. The gap between passive comprehension and active production — much larger in a second language than in a first — means that writing in the target language while listening to it at full speed produces errors that compound with fatigue.

Notes taken in the home language miss the immersive benefit. The learner is translating in real time — processing in the foreign language and transcribing in the home language — which is a different cognitive activity from sustained exposure. And the output is home-language notes that don't contribute to linguistic acquisition.

There is no clean solution within the constraint of traditional note-taking.

What AI capture provides

Record the lecture. Kuulo transcribes it in the foreign language, on-device, in real time, with no internet required. After the lecture, generate the AI summary. Review the vocabulary you didn't understand. Read the summary in the foreign language as language learning input.

This changes the relationship between lecture attendance and language acquisition:

The transcript is a reading comprehension exercise. A transcribed lecture in the target language, reviewed after the session, is reading material at exactly the level of the language you're actually encountering in the course — not simplified language-learning material, but the actual academic register of the discipline.

Vocabulary acquisition in context. The words you didn't understand in the lecture appear in the transcript in their original context. Looking up a word in the sentence you actually heard it, connected to the idea it was explaining, produces better retention than looking up words in isolation.

The summary gives you the structure. You attended the lecture but couldn't capture everything at once. The AI summary gives you the architecture of what was said — the main argument, the key examples, the conclusion. With that structure in view, the full transcript becomes easier to read, because you know where each section is going.

Live translation for total immersion

Kuulo's live translation works offline, on-device. In a language immersion context, live translation provides a different capability: the ability to follow content you're not yet competent enough to understand without support.

This is most useful in two scenarios:

The lecture where you're a step behind your linguistic competence. You're studying in Italy, your Italian is B2, and you're attending a lecture in Italian that uses domain-specific vocabulary you don't have yet. Live translation running alongside the Italian audio gives you a scaffold — you're not translating everything, but when a word or phrase is opaque, the translation is available. This is different from substituting translation for engagement; it's a linguistic scaffold that lets you engage with material that would otherwise be inaccessible.

The conversation with locals or professors that goes beyond your active vocabulary. Comprehension in a foreign language typically extends well beyond production. You may understand 90% of a native speaker's explanation and miss the crucial 10% that contains the point. A recorded and subsequently translated account fills the gap.

The Erasmus year study case

An Erasmus student in Spain, studying economics in Spanish, has a semester of lectures in a language that is not their first. The academic vocabulary of economics in Spanish — amortización, rendimiento marginal, tipos de interés — is not vocabulary most exchange students arrive with.

A typical Erasmus lecture note-taking strategy involves: missing significant content because of processing speed, writing in the home language to compensate, and reviewing notes that are incomplete and context-free.

A Kuulo-supported strategy: record the lecture in Spanish, attend and engage fully, generate the AI summary in Spanish after class, review the transcript for vocabulary gaps, and use the notes as Spanish-language reading material for exam preparation.

At the end of the Erasmus year, the student has a corpus of academic Spanish — specific to their discipline — that serves as revision material, vocabulary reference, and evidence of their linguistic development across the year.

Study abroad: language acquisition beyond the classroom

Language learning in a study abroad context happens outside lectures as much as inside them. Conversations with host families, discussions with local students, appointments at the university administration office, casual encounters that produce the kind of practical vocabulary that textbooks don't prioritise.

A 60-second voice memo into Kuulo after a significant conversation — "vocabulary from today: the landlord used 'quittance de loyer' for the rent receipt, I didn't know 'charges' means utilities in French property contexts" — generates a structured vocabulary note. After a month of daily memos, the collection is a personalised vocabulary reference built from real encounters in the country.

This is vocabulary acquisition from life, not from Duolingo. The words you learn this way are attached to actual situations, actual people, and actual problems you needed to solve — which produces significantly stronger retention than decontextualised vocabulary drilling.

The live translation workflow for multilingual situations

Study abroad situations frequently involve multiple languages. A student studying in Paris, with a Spanish flatmate and an Italian study group, operating in French in lectures, might switch between languages multiple times in a day.

Kuulo's live translation handles multiple source languages. A group study session where participants are speaking in French and Spanish can be recorded and translated into English for the student who's working beyond their current level in both. The multilingual recording isn't a problem — it's processed and translated alongside the transcription, all on-device.

For researchers conducting fieldwork in multiple languages, or graduate students working across language boundaries, this isn't a student use case — it's a practical research tool.

Privacy in language learning contexts

An exchange student recording lectures, conversations, and everyday interactions is generating audio that includes other people's voices and conversations that are sometimes personal. On-device processing means none of that audio reaches a cloud server. The recordings are on the student's device and stay there.

This matters in two directions: for the learner's own privacy (their language learning corpus, their conversations, their mistakes) and for the people they're recording (professors, local students, host family members). On-device processing applies the same principle that makes it appropriate for clinical and legal contexts to an environment where it's rarely discussed but equally relevant.

A language learner's recordings are a record of their immersive experience. Those recordings belong to the learner. On-device processing ensures they stay that way.