← WhatTNext AI सर्टिफ़िकेशन परीक्षा गाइड

सर्टिफ़ाइड AI फ़ाउंडेशनल

कोर्स और परीक्षा अभी अंग्रेज़ी में हैं। उनके आसपास के पेज हिंदी में हैं।

यह किसे प्रमाणित करता है

A Certified AI Foundational holder can explain what modern AI, and generative AI in particular, is and is not; use chat assistants effectively and safely for everyday study and work; check AI answers before relying on them; and protect their own and other people's personal data. It certifies confident, responsible everyday use of AI. It does not certify technical or specialist skills.

किसे देनी चाहिए

सुझाया गया अनुभव

None required. You should be comfortable using a smartphone or computer and reading English at about Class 10 level. Having tried a chat assistant a few times helps.

परीक्षा

  • 60 मिनट में 30 अंक वाले सवाल
  • 60% ज्ञान, 40% व्यावहारिक क्षमता
  • 3 अतिरिक्त परीक्षण सवाल जो गिने नहीं जाते, ताकि नए सवाल परखे जा सकें
  • एक जवाब, कई जवाब (सब सही हों तभी अंक) और क्रम वाले सवाल, कई असली प्रॉम्प्ट और AI जवाबों पर बने
  • नतीजा: एक तय पास मानक पर पास या फ़ेल, हर सेक्शन के लिए "मानक पूरा" या "सुधार चाहिए"
  • ऑनलाइन आँकी गई, एक बार में एक सवाल, शुरुआत में वैकल्पिक वेबकैम पहचान जाँच
कोर्स लीजिए $5 / ₹199

सेक्शन और भार

सेक्शनभारजाँचे जाने वाले कौशल
AI concepts and how generative AI works20%
  • Distinguish artificial intelligence, machine learning, deep learning and generative AI using everyday examples.
  • Explain, at an intuition level, how a large language model produces text by predicting the next token.
  • Describe what training data is and why a model's built-in knowledge stops at a cut-off date.
  • Explain why the same prompt can give different answers on different tries.
  • Identify the main kinds of generative AI output: text, images, audio, video and code.
  • Recognise what a context window is and why long conversations can lose earlier details.
Prompting and everyday use20%
  • Write a prompt that states the task, the audience, the relevant context and the format you want.
  • Improve a weak prompt by adding missing context, constraints or an example.
  • Iterate on an answer with specific follow-up instructions instead of starting again.
  • Ask an assistant to state its assumptions or list what it is unsure about.
  • Use AI to explain, summarise, translate or quiz yourself on study material in a way that supports learning.
  • Choose when giving an example of the style you want will help the assistant.
Evaluating AI output20%
  • Recognise a hallucination: a confidently stated fact, figure or citation that is not true.
  • Verify a factual claim against a primary or authoritative source.
  • Spot out-of-date information in an answer and explain why it happens.
  • Identify bias or stereotyping in an AI answer or image.
  • Check that a summary has not dropped, changed or invented key points.
  • Decide how much checking an answer needs, based on what is at stake.
Safety, privacy and the law20%
  • Identify personal and sensitive information that should not be pasted into public AI tools.
  • Describe the rights of a Data Principal under India's Digital Personal Data Protection Act 2023: information, correction and erasure, grievance redressal and nomination.
  • Explain what valid consent means under the DPDP Act 2023: free, specific, informed, unconditional and unambiguous.
  • Recognise AI-enabled scams such as cloned-voice calls and deepfake videos, and verify through a separate trusted channel.
  • Explain why AI-generated content should be disclosed and labelled.
  • Respect copyright and academic-integrity rules when using AI for assignments and creative work.
Choosing tools and knowing when not to use AI10%
  • Match an everyday task to a suitable tool: chat assistant, search engine, calculator or spreadsheet, image generator, or no AI at all.
  • Explain why a chat assistant is a poor choice for exact calculations or live information unless it uses a calculator or search tool.
  • Recognise situations where AI should not be used, such as banned graded work, decisions about people or emergencies.
  • Compare what AI does well and badly for writing, research, studying and images.
AI, work and careers10%
  • Distinguish AI automating a task from AI augmenting a person's role.
  • Identify which tasks in a given job are more and less exposed to generative AI.
  • Describe skills that tend to grow in value as AI spreads, such as judgement, communication and domain knowledge.
  • Plan one concrete way to use AI to learn a skill for your career, with a check that you really learned it.

यह स्तर किन ढाँचों के अनुरूप है

हमने पाठ्यक्रम बनाने में इन सार्वजनिक ढाँचों का इस्तेमाल किया। ये इस सर्टिफ़िकेशन का समर्थन नहीं करते।

नमूना सवाल

ये शैली और स्तर के उदाहरण हैं। ये परीक्षा में नहीं आते।

1.

Which statement best describes how a large language model (LLM) such as the one behind a chat assistant produces its reply?

  1. It looks up the reply in a database of facts that human editors checked and approved during training.
  2. It generates the reply piece by piece, each time predicting a likely next token from patterns learned in training.
  3. It finds the most popular web page on the topic and rewrites that page in simpler language for you.
  4. It follows a fixed decision tree that programmers wrote by hand in advance for every type of question people might ask.

जवाब: B

LLMs generate text token by token by predicting what is likely to come next, based on patterns learned from training data. They do not consult a verified fact database, which is why they can produce fluent but false statements.

2.

A chat assistant with no web search tells you a tournament "is scheduled for next month". In fact it was played and finished last year. What is the most likely reason?

  1. Its servers run in a different time zone from yours, so every date it reports is shifted by about a year.
  2. It is programmed to avoid giving results of sporting events.
  3. Its training data ends before the tournament took place, and it cannot check current information.
  4. The assistant deliberately gives old answers to save computing power.

जवाब: C

A model's built-in knowledge stops at its training cut-off. Without a search tool it cannot know about later events, yet it may still answer confidently.

3.

Meera is revising photosynthesis for her Class 12 board exam. She wants to practise, not just read. Which prompt is most likely to help her learn?

  1. "Write a very detailed and complete essay on photosynthesis covering every concept, diagram and possible exam question so I can memorise all of it tonight before the exam."
  2. "Photosynthesis questions."
  3. "I'm a Class 12 student revising photosynthesis. Ask me 5 questions one at a time, easy to hard. Wait for my answer, then tell me what I got right or wrong."
  4. "Give me the exact questions that will come in this year's board exam on photosynthesis, with answers."

जवाब: C

The strongest prompt states who she is, the goal (practice), the format (one question at a time, rising difficulty) and the feedback she wants. Asking for future exam questions invites made-up content, and a long essay doesn't make her practise.

4.

A student used a chat assistant for a civics project. Read the prompt and the answer. Which statement in the answer is wrong and must be corrected before it is used?

प्रॉम्प्टGive me three key facts about India's Digital Personal Data Protection Act.
AI का जवाब1. The Act was passed by Parliament in 2023. 2. A Data Fiduciary that fails to take reasonable security safeguards to prevent a personal data breach can face a penalty of up to ₹250 crore. 3. The Act requires every company to store all personal data only on servers located inside India.
  1. Statement 1, because the Act was passed in 2019.
  2. Statement 2, because the maximum penalty is ₹25 lakh.
  3. Statement 3, because the Act does not require all data to stay in India.
  4. Statement 2, because these penalties apply only to government bodies and not to private companies.

जवाब: C

Statements 1 and 2 match the Act (assented 11 August 2023; Schedule item 1: up to ₹250 crore). The Act does not impose a blanket data-localisation rule; section 16 lets the government restrict transfers to countries it notifies. The answer mixes true and false facts in the same confident tone, which is typical of hallucination.

5.

Under India's Digital Personal Data Protection Act 2023, who counts as a "child"?

  1. Anyone who has not completed 13 years of age
  2. Anyone who has not completed 16 years of age
  3. Anyone who has not completed 18 years of age
  4. Anyone who has not completed 21 years of age

जवाब: C

Section 2(f) defines a child as an individual who has not completed eighteen years. Processing a child's data needs verifiable consent of a parent or lawful guardian (section 9).

6.

You receive a WhatsApp voice note that sounds exactly like your cousin. She is crying and asks you to send ₹40,000 by UPI to a new number "right now" because of an accident. What should you do first?

  1. Send a small amount first, such as ₹1,000, to test whether the number is genuine, and then send the rest once it goes through.
  2. Call your cousin, or a family member who is with her, on a number you already have saved, before sending anything.
  3. Reply asking her to send a selfie video, since a video would prove it is really her.
  4. Upload the voice note to a free online deepfake detector and pay if it says the voice is real.

जवाब: B

Voices and videos can now be cloned convincingly, and online detectors are unreliable. The safe step is to verify through a separate channel you already trust before paying anything. A 'test' payment still sends money to a possible scammer.

7.

Which of these tasks is the best fit for a general chat assistant used on its own, with no web search, calculator or file tools switched on?

  1. Working out the exact EMI, to the rupee, for 37 different loans with different interest rates and tenures
  2. Brainstorming ten title ideas for your college fest poster
  3. Checking whether your train is running late today
  4. Shortlisting job applicants by looking at their photographs

जवाब: B

Idea generation suits a chat assistant. Exact repeated calculations belong in a spreadsheet or calculator, live status needs a live source, and judging people by photos is unfair and should not be done with or without AI.

8.

Which example shows AI augmenting a person's role rather than automating a task completely?

  1. A support chatbot answers every customer query from start to finish, with no human ever involved
  2. Software reads supplier invoices and posts them to the accounts with no review
  3. A radiologist reviews areas an AI tool has flagged, then decides and signs the report
  4. A robot arm takes over one step on a factory assembly line from a worker

जवाब: C

Augmentation means the person keeps the judgement and responsibility while AI helps. In the other options the task is done end to end by the machine.