The short answer
"AI" stands for artificial intelligence. In Arabic it is «الذكاء الاصطناعي» (al-dhakāʾ al-iṣṭināʿī).
- The full name is pronounced roughly "ar-ti-FISH-ul in-TEL-i-jens". Cambridge Dictionary gives the British pronunciation as /ˌɑː.tɪ.fɪʃ.əl ɪnˈtel.ɪ.dʒəns/ (Cambridge Dictionary).
- The abbreviation AI is said letter by letter: "ay-eye", /ˌeɪˈaɪ/ (Cambridge Dictionary; Oxford Learner's Dictionaries).
"Artificial" is «اصطناعي» and "intelligence" is «ذكاء». English puts the adjective before the noun, so the two words come in the opposite order to the Arabic.
The English name has been in use since the 1950s. It appears in the title of a research proposal dated 31 August 1955, written by John McCarthy and three colleagues, which proposed a summer study of artificial intelligence at Dartmouth College in the United States in the summer of 1956 (McCarthy et al., reprinted in AI Magazine, 2006).
How to use this page
We have gathered 30 terms that non-experts often meet in the news, at work and in chat apps. Each term has its English name, its Arabic name and a one-sentence explanation. The English name links to the term's entry in our glossary, where you will find a fuller definition and articles to read next.
The basics
| English | Arabic | In one sentence |
|---|---|---|
| Artificial intelligence (AI) | الذكاء الاصطناعي | Computer software that does tasks we usually link with human thinking, such as understanding speech, translating, summarising or spotting patterns. |
| Algorithm | الخوارزمية | A set of step-by-step instructions a computer follows to solve a problem. |
| Machine learning | تعلّم الآلة (التعلّم الآلي) | A way of building AI in which the computer learns patterns from many examples instead of being given fixed rules. |
| Neural network / deep learning | الشبكة العصبية الاصطناعية / التعلّم العميق | A machine-learning system loosely inspired by the brain; "deep" means it has many layers, and it is the approach behind most of today's AI. |
| Model (AI model) | النموذج (نموذج الذكاء الاصطناعي) | The trained "engine" inside an AI system: it takes an input, such as a question or an image, and produces an output, such as an answer or a label. |
| Training | التدريب | The stage in which a model learns from large amounts of data, usually over weeks or months. |
| Training data | بيانات التدريب | The text, images, audio or other material a model learns from; gaps in it, such as too little Arabic content, show up later as weaknesses or bias. |
| Parameters | المُعامِلات | The internal numbers a model adjusts during training, often counted in billions. |
| Inference | تشغيل النموذج (الاستدلال) | Using a model after it has been trained: every question you ask a chatbot is inference. |
The tools you use
| English | Arabic | In one sentence |
|---|---|---|
| Generative AI | الذكاء الاصطناعي التوليدي | AI that creates new content on request, such as text, images, audio or code. |
| Large language model (LLM) | النموذج اللغوي الكبير | A model trained on very large amounts of text so it can read and write language; most chat assistants run on one. |
| Arabic LLM | النموذج اللغوي العربي الكبير | A large language model built with Arabic as a main language from the start; Gulf examples include Jais, ALLaM and Fanar. |
| Chatbot / AI assistant | روبوت المحادثة / المساعد الذكي | An app you talk to in ordinary language, by typing or speaking, which answers using an AI model. |
| Prompt | الموجِّه (نص الطلب) | The question or instruction you give an AI tool; a clear prompt with context usually gets a better answer. |
| Multimodal | متعدد الوسائط | Describes AI that can work with more than one kind of input or output, such as text and images together. |
| Reasoning model | نموذج استدلالي | A model that works through a problem in steps before answering, so it is often better at maths, logic and coding, but slower and more costly. |
| AI agent | وكيل الذكاء الاصطناعي | An AI tool that doesn't just answer but carries out several steps towards a goal, ideally with a person approving what matters. |
| Agentic AI | الذكاء الاصطناعي الوكيل | AI systems built from agents that can carry out a sequence of tasks on their own within set limits. |
Trust and safety
| English | Arabic | In one sentence |
|---|---|---|
| Hallucination | الهلوسة (الاختلاق) | When an AI tool states something false as if it were true, which is why important answers are worth checking. |
| Algorithmic bias | الانحياز الخوارزمي | When an AI system treats some groups unfairly or reflects stereotypes, usually because of gaps in its training data. |
| Automation bias | الانحياز للأتمتة | The human habit of trusting a machine's answer too readily, even when it is wrong. |
| Deepfake | التزييف العميق | A fake video, image or voice recording made with AI to look or sound like a real person. |
| Human-in-the-loop | الإنسان داخل دائرة القرار | A design in which AI suggests, and a person approves or rejects before anything important happens. |
| Benchmark | الاختبار المعياري | A standard test used to compare AI models; useful, but a good score does not guarantee a model does well on real tasks. |
Infrastructure
| English | Arabic | In one sentence |
|---|---|---|
| Compute | القدرة الحاسوبية | The processing power used to train and run AI, mostly from specialised chips in data centres. |
| GPU (graphics processing unit) | وحدة معالجة الرسومات | A chip first designed for video-game graphics that turned out to be very good at the maths AI needs. |
| Data centre | مركز البيانات | A large building full of computers, storage and cooling equipment where models are trained and run; it needs a lot of reliable electricity. |
The big debates
| English | Arabic | In one sentence |
|---|---|---|
| AGI (artificial general intelligence) | الذكاء الاصطناعي العام | A loose term for AI that could handle most intellectual tasks about as well as a capable person; there is no agreed definition. |
| ASI (artificial superintelligence) | الذكاء الاصطناعي الفائق | A hypothetical AI far more capable than humans in almost every area; it does not exist today. |
| AI safety | سلامة الذكاء الاصطناعي | The field that studies how to prevent AI from causing harm, through testing, safeguards, oversight and rules. |
Terms you may meet in more than one Arabic form
Arabic translations of some terms have not settled yet, so you may see the same idea under more than one name:
- Machine learning: «تعلّم الآلة» and «التعلّم الآلي» are two names for the same thing.
- Hallucination: «الهلوسة» is the most widespread, and some editors prefer «الاختلاق» ("fabrication") because it is clearer.
- Prompt: «الموجِّه» or «نص الطلب» ("request text").
- Inference: often translated «الاستدلال», but we use «تشغيل النموذج» ("running the model") so it is not confused with «النموذج الاستدلالي», the reasoning model, which is something different.
Further reading
- Our full glossary, with the terms on this page and more, and fuller definitions in English and Arabic.
- The Data and AI Lexicon («معجم البيانات والذكاء الاصطناعي»), launched in 2022 by the Saudi Data and AI Authority (SDAIA) with the King Salman Global Academy for Arabic Language. It gives short definitions of terms in Arabic and English (Ajel, 25 March 2022; Saudipedia).