{"course":{"slug":"ai-for-students","title":"AI for Students","url":"https://addaly.com/learn/ai-for-students","level":"beginner","track":"work"},"lesson":{"slug":"learning-a-language","title":"Languages, where it is genuinely excellent","url":"https://addaly.com/learn/ai-for-students/learning-a-language","position":35,"of":73,"minutes":9,"answer":"The tool is genuinely excellent for input, explanation and drilling, but an agreeable partner that understands your errors without flagging them will fossilise them: demand a correction list after every message.","keyIdea":"The tool is genuinely excellent for input, explanation and drilling, but an agreeable partner that understands your errors without flagging them will fossilise them: demand a correction list after every message."},"body":{"format":"markdown","text":"## The clearest win in this course\n\nLanguage learning is where these tools deliver the most, and it is worth saying plainly after several lessons of caution. A patient conversation partner, available at any hour, who will speak at your level, never gets bored, never judges your accent, and will explain the grammar of the sentence you just got wrong. This did not exist before, at any price, for most learners.\n\nThe things that work:\n\n- **Conversation at a controlled level.** *Talk to me in Spanish at roughly A2. Correct me only at the end of each of my messages, in English, in one line.*\n- **Comprehensible input on demand.** A short text about a topic you care about, at your level, is the single best thing for acquisition, and it has always been the hardest thing to find. Now you can generate it endlessly.\n- **Explaining why**, which textbooks compress. Why the subjunctive here and not there. Why this particle. Why the word order changed.\n- **Drilling on the pattern you personally keep getting wrong**, rather than a generic exercise book.\n- **Reading support.** Paste a paragraph, get a gloss, then re-read the original.\n\n## Where it fails, and the failure is specific\n\n**Low-resource languages.** Quality tracks the amount of training text. Spanish, French, German, Mandarin, Japanese: strong. Many South Asian, African, indigenous and minority languages: much weaker, sometimes confidently wrong on basic grammar, and prone to producing text that is fluent-sounding and not idiomatic. If you are learning a language with a small web presence, verify with a human speaker or a real textbook. This is one of the sharper inequalities in the technology and it is not improving as fast as the headline capabilities.\n\n**Register and dialect.** Models default to a standard formal register. They will happily produce textbook Arabic to someone who needs Egyptian, or a formality level that would be strange between friends. Ask explicitly for the variety and register you need, and be aware it may not have it.\n\n**Recency in slang and usage.** By definition out of date.\n\n**Pronunciation.** Text feedback on pronunciation is close to useless. Use audio: **Forvo** for individual words by native speakers, and speech tools that give real audio. Many models now support voice conversation, which helps considerably.\n\n## Fossilisation: the risk worth naming\n\nIn second-language research, **fossilisation** is an error that becomes permanent: you have said it wrong so many times that it now feels correct. It happens to learners who produce a great deal without correction.\n\nAn AI conversation partner produces exactly that condition unless you set it up otherwise, because assistants are tuned to be agreeable and will often understand and continue past an error rather than flag it. You will have a pleasant fluent conversation full of mistakes nobody mentioned.\n\nThe fix is one clause, used every time:\n\n> After each of my messages, before replying, list any errors I made and the correct form. Keep the list short. Then continue the conversation naturally.\n\nAnd once a week, ask: *what mistakes have I made repeatedly in this conversation?* Those are the ones about to fossilise, and they are the ones worth drilling.\n\n## Translation, and what it costs you\n\nTranslating your thoughts from your first language into the target one is how you build production ability. Machine translation removes that work entirely, which is why heavy translation use produces learners who can read comfortably and cannot speak.\n\nA defensible arrangement: write it yourself first, then compare with a translation, then note the differences. The comparison is where the learning is. Going straight to the translation is where it is not.\n\nThe honest exception: when translation is not the goal (you need to read a paper in another language for your dissertation, or communicate something urgent) use the tool and do not feel bad about it. **DeepL** and Google Translate both have free tiers and are excellent for major language pairs.\n\n## Free tools worth the space\n\n- **Anki** for vocabulary, with your own cards, and audio pulled from Forvo.\n- **Language Reactor** or similar for subtitles, for input from material you actually want to watch.\n- **Tatoeba**, a free database of example sentences with translations, human-produced.\n- **Wiktionary**, which is far better than its reputation for etymology and inflection tables.\n- A human. Language exchange partners cost nothing and provide the one thing no model does: a person whose good opinion you want, who will be confused when you are unclear, and whose confusion is real.\n\n## For students studying in a second language\n\nA large share of students worldwide study in a language that is not their first. Using a model to check your English, understand a lecture, or rephrase a confusing question is entirely legitimate and levels a genuine unfairness.\n\nTwo cautions. First, some institutions treat AI-assisted language editing as needing disclosure; check your policy and disclose if in doubt, which costs nothing. Second, detectors flag non-native writing at higher rates than native writing: a well-documented and serious problem covered in Module 7. Keep your drafts and version history. That is the protection, and you should have it whether or not you use any tool."},"check":{"question":"A learner has a fluent 40-minute conversation in French with an assistant and receives no corrections. Why is this a problem rather than a sign of progress?","options":[{"text":"Because conversation without a native speaker cannot improve production, only comprehension.","correct":false,"tempting":"Denies a benefit the lesson explicitly credits; controlled-level conversation does build production, provided errors are surfaced."},{"text":"Because the model was probably operating above the learner's level, so the fluency was the model's rather than theirs.","correct":false,"tempting":"Misplaces the fault in level setting; the described problem persists at a perfectly matched level."},{"text":"Because uncorrected repeated errors can fossilise, and an agreeable partner will understand and continue past them.","correct":true},{"text":"Because forty minutes of unstructured conversation is less efficient than the same time spent on spaced vocabulary drilling.","correct":false,"tempting":"Swaps a diagnosis for a preference about time allocation, and understates the role of extended input in acquisition."}]},"prev":{"slug":"literature-and-close-reading","title":"Literature, and the text it half remembers","url":"https://addaly.com/learn/ai-for-students/literature-and-close-reading"},"next":{"slug":"statistics-for-coursework","title":"Statistics for coursework","url":"https://addaly.com/learn/ai-for-students/statistics-for-coursework"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Languages, where it is genuinely excellent', in AI for Students. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-for-students/learning-a-language (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Languages, where it is genuinely excellent","courseTitle":"AI for Students","module":{"index":4,"slug":"subject-by-subject","title":"Subject by subject"},"answer":null,"keyIdea":"The tool is genuinely excellent for input, explanation and drilling, but an agreeable partner that understands your errors without flagging them will fossilise them: demand a correction list after every message.","hero":{"kind":"plate","image":null,"alt":null,"caption":null},"blocks":[{"t":"h2","id":"the-clearest-win-in-this-course","runs":[{"text":"The clearest win in this course"}]},{"t":"p","runs":[{"text":"Language learning is where these tools deliver the most, and it is worth saying plainly after several lessons of caution. A patient conversation partner, available at any hour, who will speak at your level, never gets bored, never judges your accent, and will explain the grammar of the sentence you just got wrong. This did not exist before, at any price, for most learners."}]},{"t":"p","runs":[{"text":"The things that work:"}]},{"t":"ul","items":[[{"text":"Conversation at a controlled level.","b":true},{"text":" "},{"text":"Talk to me in Spanish at roughly A2. Correct me only at the end of each of my messages, in English, in one line.","i":true}],[{"text":"Comprehensible input on demand.","b":true},{"text":" A short text about a topic you care about, at your level, is the single best thing for acquisition, and it has always been the hardest thing to find. Now you can generate it endlessly."}],[{"text":"Explaining why","b":true},{"text":", which textbooks compress. Why the subjunctive here and not there. Why this particle. Why the word order changed."}],[{"text":"Drilling on the pattern you personally keep getting wrong","b":true},{"text":", rather than a generic exercise book."}],[{"text":"Reading support.","b":true},{"text":" Paste a paragraph, get a gloss, then re-read the original."}]]},{"t":"h2","id":"where-it-fails-and-the-failure-is-specific","runs":[{"text":"Where it fails, and the failure is specific"}]},{"t":"p","runs":[{"text":"Low-resource languages.","b":true},{"text":" Quality tracks the amount of training text. Spanish, French, German, Mandarin, Japanese: strong. Many South Asian, African, indigenous and minority languages: much weaker, sometimes confidently wrong on basic grammar, and prone to producing text that is fluent-sounding and not idiomatic. If you are learning a language with a small web presence, verify with a human speaker or a real textbook. This is one of the sharper inequalities in the technology and it is not improving as fast as the headline capabilities."}]},{"t":"p","runs":[{"text":"Register and dialect.","b":true},{"text":" Models default to a standard formal register. They will happily produce textbook Arabic to someone who needs Egyptian, or a formality level that would be strange between friends. Ask explicitly for the variety and register you need, and be aware it may not have it."}]},{"t":"p","runs":[{"text":"Recency in slang and usage.","b":true},{"text":" By definition out of date."}]},{"t":"p","runs":[{"text":"Pronunciation.","b":true},{"text":" Text feedback on pronunciation is close to useless. Use audio: "},{"text":"Forvo","b":true},{"text":" for individual words by native speakers, and speech tools that give real audio. Many models now support voice conversation, which helps considerably."}]},{"t":"h2","id":"fossilisation-the-risk-worth-naming","runs":[{"text":"Fossilisation: the risk worth naming"}]},{"t":"p","runs":[{"text":"In second-language research, "},{"text":"fossilisation","b":true},{"text":" is an error that becomes permanent: you have said it wrong so many times that it now feels correct. It happens to learners who produce a great deal without correction."}]},{"t":"p","runs":[{"text":"An AI conversation partner produces exactly that condition unless you set it up otherwise, because assistants are tuned to be agreeable and will often understand and continue past an error rather than flag it. You will have a pleasant fluent conversation full of mistakes nobody mentioned."}]},{"t":"p","runs":[{"text":"The fix is one clause, used every time:"}]},{"t":"quote","runs":[{"text":"After each of my messages, before replying, list any errors I made and the correct form. Keep the list short. Then continue the conversation naturally."}]},{"t":"p","runs":[{"text":"And once a week, ask: "},{"text":"what mistakes have I made repeatedly in this conversation?","i":true},{"text":" Those are the ones about to fossilise, and they are the ones worth drilling."}]},{"t":"h2","id":"translation-and-what-it-costs-you","runs":[{"text":"Translation, and what it costs you"}]},{"t":"p","runs":[{"text":"Translating your thoughts from your first language into the target one is how you build production ability. Machine translation removes that work entirely, which is why heavy translation use produces learners who can read comfortably and cannot speak."}]},{"t":"p","runs":[{"text":"A defensible arrangement: write it yourself first, then compare with a translation, then note the differences. The comparison is where the learning is. Going straight to the translation is where it is not."}]},{"t":"p","runs":[{"text":"The honest exception: when translation is not the goal (you need to read a paper in another language for your dissertation, or communicate something urgent) use the tool and do not feel bad about it. "},{"text":"DeepL","b":true},{"text":" and Google Translate both have free tiers and are excellent for major language pairs."}]},{"t":"h2","id":"free-tools-worth-the-space","runs":[{"text":"Free tools worth the space"}]},{"t":"ul","items":[[{"text":"Anki","b":true},{"text":" for vocabulary, with your own cards, and audio pulled from Forvo."}],[{"text":"Language Reactor","b":true},{"text":" or similar for subtitles, for input from material you actually want to watch."}],[{"text":"Tatoeba","b":true},{"text":", a free database of example sentences with translations, human-produced."}],[{"text":"Wiktionary","b":true},{"text":", which is far better than its reputation for etymology and inflection tables."}],[{"text":"A human. Language exchange partners cost nothing and provide the one thing no model does: a person whose good opinion you want, who will be confused when you are unclear, and whose confusion is real."}]]},{"t":"h2","id":"for-students-studying-in-a-second-language","runs":[{"text":"For students studying in a second language"}]},{"t":"p","runs":[{"text":"A large share of students worldwide study in a language that is not their first. Using a model to check your English, understand a lecture, or rephrase a confusing question is entirely legitimate and levels a genuine unfairness."}]},{"t":"p","runs":[{"text":"Two cautions. First, some institutions treat AI-assisted language editing as needing disclosure; check your policy and disclose if in doubt, which costs nothing. Second, detectors flag non-native writing at higher rates than native writing: a well-documented and serious problem covered in Module 7. Keep your drafts and version history. That is the protection, and you should have it whether or not you use any tool."}]}],"check":{"question":[{"text":"A learner has a fluent 40-minute conversation in French with an assistant and receives no corrections. Why is this a problem rather than a sign of progress?"}],"options":[{"text":[{"text":"Because conversation without a native speaker cannot improve production, only comprehension."}],"correct":false,"tempting":[{"text":"Denies a benefit the lesson explicitly credits; controlled-level conversation does build production, provided errors are surfaced."}]},{"text":[{"text":"Because the model was probably operating above the learner’s level, so the fluency was the model’s rather than theirs."}],"correct":false,"tempting":[{"text":"Misplaces the fault in level setting; the described problem persists at a perfectly matched level."}]},{"text":[{"text":"Because uncorrected repeated errors can fossilise, and an agreeable partner will understand and continue past them."}],"correct":true,"tempting":null},{"text":[{"text":"Because forty minutes of unstructured conversation is less efficient than the same time spent on spaced vocabulary drilling."}],"correct":false,"tempting":[{"text":"Swaps a diagnosis for a preference about time allocation, and understates the role of extended input in acquisition."}]}],"order":[0,2,3,1],"why":null},"prev":{"slug":"literature-and-close-reading","title":"Literature, and the text it half remembers","minutes":8},"next":{"slug":"statistics-for-coursework","title":"Statistics for coursework","minutes":9}}}