Sun’iy intellekt inglizcha artificial intelligence (AI) — “artificial intellect” emas. Eng ko‘p kerak bo‘ladigan beshta atama: model (o‘rgatilgan dastur), prompt (AI’ga yozgan ko‘rsatmangiz), dataset (ma’lumotlar to‘plami), machine learning (mashinaviy o‘qitish — dastur misollardan o‘rganadi) va hallucination (AI to‘qib chiqargan noto‘g‘ri javob). Quyida jami 64 ta atama to‘rt guruhga bo‘lingan. Har birida o‘zbekcha tarjima va misol gap bor.
Atamalar oddiy tilda nimani anglatishi sun’iy intellekt nima sahifasida tushuntirilgan. Bu sahifa — lug‘at: inglizcha maqola o‘qiganda, AI haqida gapirganda yoki IT sohasida ishlaganda kerak bo‘ladigan so‘zlar.








1. Asosiy tushunchalar (17 ta)
| Inglizcha | O‘zbekcha | Misol |
|---|---|---|
| artificial intelligence (AI) | sun’iy intellekt | AI helps doctors read X-rays faster. |
| machine learning | mashinaviy o‘qitish | Machine learning lets a program improve from examples. |
| deep learning | chuqur o‘qitish | Deep learning is used in speech recognition. |
| neural network | neyron tarmoq | A neural network has many layers of simple units. |
| algorithm | algoritm | The app uses an algorithm to recommend videos. |
| model | model (o‘rgatilgan dastur) | This model can answer questions in Uzbek. |
| data | ma’lumotlar | The data comes from public websites. |
| dataset | ma’lumotlar to‘plami | The team built a dataset of Uzbek speech recordings. |
| training | o‘qitish (modelni) | Training a large model can take weeks. |
| training data | o‘qitish ma’lumotlari | Poor training data leads to poor answers. |
| parameter | parametr | The model has billions of parameters. |
| generative AI | generativ AI (yangi matn, rasm yaratadigan) | Generative AI can draw a picture from a short description. |
| large language model (LLM) | katta til modeli | An LLM predicts the next word in a sentence. |
| natural language processing (NLP) | tabiiy tilni qayta ishlash | NLP helps computers understand what people write. |
| computer vision | kompyuter ko‘rishi | Computer vision recognises number plates on the road. |
| speech recognition | nutqni aniqlash | Speech recognition turns your voice into text. |
| text-to-speech | matnni ovozga aylantirish | Text-to-speech reads the answer aloud. |
2. Chatbot va AI vositalari bilan ishlash (18 ta)
| Inglizcha | O‘zbekcha | Misol |
|---|---|---|
| prompt | promt, so‘rov-ko‘rsatma | Write a clear prompt and mention your level. |
| prompt engineering | promt tuzish mahorati | Prompt engineering is now a useful job skill. |
| chatbot | chatbot, suhbat boti | The bank’s chatbot answers simple questions. |
| AI assistant | AI yordamchi | I asked my AI assistant to plan my study week. |
| input | kiritilgan ma’lumot | The input was a photo of a restaurant menu. |
| output | natija | Always check the output before you use it. |
| response | javob | The response was too long, so I asked for a shorter one. |
| context window | kontekst oynasi (model bir vaqtda “ko‘radigan” matn hajmi) | A very long book may not fit in the context window. |
| token | token (matnning kichik bo‘lagi) | Long texts use more tokens. |
| custom instructions | shaxsiy ko‘rsatmalar | I set custom instructions to reply in simple English. |
| system prompt | tizim ko‘rsatmasi | The system prompt tells the bot how to behave. |
| regenerate | qayta yaratish | Click regenerate if you don’t like the answer. |
| upload | yuklash | You can upload a PDF and ask questions about it. |
| voice mode | ovozli rejim | I practise speaking in voice mode every evening. |
| multimodal | multimodal (matn, rasm, ovoz bilan ishlaydigan) | A multimodal model understands both text and images. |
| fine-tuning | qo‘shimcha o‘qitish, moslashtirish | Fine-tuning adapts a model to legal documents. |
| AI agent | AI agent (bir necha qadamni o‘zi bajaradigan) | An AI agent can compare prices on several websites. |
| API | API (dasturlararo interfeys) | Developers connect their apps to the model through an API. |
Promtni qanday tuzish kerakligi — ya’ni rol, vazifa va format — promt yozishni o‘rganish sahifasida.
3. Sifat, xavf va etika (17 ta)
| Inglizcha | O‘zbekcha | Misol |
|---|---|---|
| hallucination | gallyutsinatsiya (to‘qima javob) | The book title it gave me was a hallucination. |
| accuracy | aniqlik | Accuracy is often lower for less common languages. |
| bias | tarafkashlik | Bias in the data can lead to unfair decisions. |
| fact-check | faktni tekshirmoq | Always fact-check dates and numbers. |
| source | manba | Ask the chatbot for its source, then open it yourself. |
| citation | iqtibos, manbaga havola | The citation led to a page that does not exist. |
| misinformation | noto‘g‘ri ma’lumot | AI can produce misinformation that looks very convincing. |
| deepfake | dipfeyk (soxta video yoki ovoz) | The video of the famous singer was a deepfake. |
| privacy | maxfiylik | Read the privacy policy before you upload files. |
| personal data | shaxsiy ma’lumotlar | Never share your passport number with a chatbot. |
| consent | rozilik | Record someone’s voice only with their consent. |
| transparency | shaffoflik | Transparency means explaining how the system makes decisions. |
| human oversight | inson nazorati | Medical AI tools need human oversight. |
| accountability | javobgarlik | Who has accountability when an AI system makes a mistake? |
| copyright | mualliflik huquqi | Check the copyright before you use a picture in your project. |
| plagiarism | plagiat, o‘g‘irlangan matn | Submitting an AI-written essay as your own may count as plagiarism. |
| academic integrity | akademik halollik | Our university has clear rules on academic integrity. |








AI qanday xato qilishi va ularni qanday ushlash kerakligi AI xatolari va gallyutsinatsiya sahifasida.
4. Ish va texnologiya (12 ta)
| Inglizcha | O‘zbekcha | Misol |
|---|---|---|
| data scientist | ma’lumotlar tahlilchisi | She works as a data scientist at a bank in Tashkent. |
| machine learning engineer | mashinaviy o‘qitish muhandisi | Machine learning engineers put models into real products. |
| GPU | grafik protsessor | Training needs a lot of powerful GPUs. |
| cloud | bulut (bulutli xizmat) | The model runs in the cloud, not on your phone. |
| open-source | ochiq kodli | Some open-source models can be downloaded for free. |
| benchmark | benchmark, taqqoslash testi | The new model scored well on a maths benchmark. |
| inference | xulosa chiqarish (tayyor modelni ishlatish) | Inference is much cheaper than training. |
| deploy | ishga tushirmoq, joriy qilmoq | They deployed the chatbot on their website last month. |
| automation | avtomatlashtirish | Automation saves time on routine office tasks. |
| robotics | robototexnika | Robotics combines AI with mechanical engineering. |
| annotation | belgilash, annotatsiya | Annotation of images is often done by people. |
| scalable | kengaytiriladigan | A scalable service works for ten users or ten million. |






Dasturchilar uchun alohida lug‘at (bug, deploy, merge, variable) dasturlash so‘zlari sahifasida.
O‘zbek o‘quvchilari ko‘p qiladigan xatolar
- **“Artificial intellect” emas, artificial intelligence.** Ruscha “искусственный интеллект” va o‘zbekcha “sun’iy intellekt” ta’sirida chiqadigan eng mashhur xato.
- **Modelni train qilishadi, learn emas.** O‘zbekcha “o‘qitmoq” — train, “o‘rganmoq” — learn. To‘g‘ri: They trained the model on Uzbek texts. Model o‘zi esa learns from data.
- **Artikl: an AI model.** AI harf bilan o‘qiladi — /eɪ/ unli tovush bilan boshlanadi, shuning uchun an: an AI tool, an API. Lekin a GPU, chunki /dʒiː/ undosh bilan boshlanadi.
- ***Data* bilan fe’l.** Kundalik va IT tilida The data is ready keng tarqalgan; ilmiy matnda The data are… ham uchraydi. Ikkalasi ham to‘g‘ri, lekin bitta matnda bittasini ishlating. Datas esa xato.
- **Promtni yozasiz, qilmaysiz.** Make a prompt emas, write a prompt yoki give a prompt.
Talaffuzi qiyin atamalar
| So‘z | Talaffuz | Izoh |
|---|---|---|
| algorithm | /ˈælɡərɪðəm/ | Urg‘u birinchi bo‘g‘inda; th — /ð/, “s” emas |
| AI | /ˌeɪ ˈaɪ/ | Harflar bilan: “ey-ay”, “ai” emas |
| data | /ˈdeɪtə/ yoki /ˈdɑːtə/ | Ikkalasi ham to‘g‘ri |
| neural | /ˈnjʊərəl/ (UK), /ˈnʊrəl/ (US) | “neyral” emas |
| bias | /ˈbaɪəs/ | “bays” deb boshlanadi |
| parameter | /pəˈræmɪtə/ | Urg‘u ikkinchi bo‘g‘inda: pa-RA-meter |
| GPU | /ˌdʒiː piː ˈjuː/ | Harflar bilan |
| inference | /ˈɪnfərəns/ | Urg‘u birinchi bo‘g‘inda |
Talaffuzni lug‘atdagi audio bilan tekshiring: Cambridge yoki Oxford Learner’s lug‘atida ikkala aksent ham bor.
Mashq: bo‘sh joyni to‘ldiring
So‘zlar: prompt, hallucination, dataset, train, bias, deepfake, fact-check, upload, privacy, accuracy
- 1. The chatbot gave me a quote from a book that doesn’t exist. That was a ________.
- 2. Before I ________ my CV, I removed my phone number and address.
- 3. My ________ was too short, so the answer was too general.
- 4. They collected 10,000 recordings to create a ________ of Uzbek speech.
- 5. It takes a lot of time and money to ________ a large model.
- 6. The system rejected more women than men for the job. That shows ________.
- 7. The video looked real, but it was a ________.
- 8. Always ________ numbers and dates before you put them in your essay.
- 9. The app’s ________ policy explains what happens to your voice recordings.
- 10. Speech recognition has lower ________ in noisy places.
Javoblar
- 1. hallucination
- 2. upload
- 3. prompt
- 4. dataset
- 5. train
- 6. bias
- 7. deepfake
- 8. fact-check
- 9. privacy
- 10. accuracy
8 va undan ko‘p to‘g‘ri javob — yaxshi natija. Xato qilgan so‘zlaringiz uchun o‘zingiz haqida bittadan gap tuzing va ovoz chiqarib ayting. AI sohasida ishlamoqchi bo‘lsangiz, kerakli ingliz tili darajasi va o‘rganish yo‘li AI sohasida ishlash uchun ingliz tili sahifasida.
BandMate bilan
Yangi atamalarni gapda ishlatganingizdagina ular esda qoladi. BandMateda Bandi bilan texnologiya mavzusida ovozli suhbat qiling: How do you use AI at work or school? kabi savolga javob berib, shu sahifadagi 5–6 atamani ishlating. Xatolaringiz o‘zbekcha tushuntiriladi; talaffuz mashqida esa algorithm va parameter kabi so‘zlarni takrorlashingiz mumkin.
Bugun sinab ko‘ring: jadvallardan o‘zingizga eng kerakli 10 ta atamani tanlang va har biri bilan ishingiz yoki o‘qishingiz haqida bitta gap yozing.



