Generative AI first steps: Generative AI and chat assistants
Pick an option and check it. The explanation shows after each answer. Take as long as you like; nothing is timed and nothing is sent.
15 questions
Question 1 · easy
Asha is a clerk at a bank in Sambalpur. Four AI features help her today. Which one is generating?
Ask whether the output was already in what went in. The farewell speech did not exist until the feature wrote it, so it is generated. Reading the amount on a cheque, checking a signature and putting letters into piles all read, check or sort what is already there, which is recognising.
Question 2 · easy
In the name large language model, what does the word language tell you?
Language means the model works with words: it learned patterns from a huge amount of text and uses them to write new text. The word does not mean it works only in English, and translating is not the only thing it does.
Question 3 · easy
Meena, a teacher in Berhampur, types one line: "a slow, gentle tune for a lullaby in our Class 1 play". Back comes new music that nobody played. Which of the four kinds of things generative AI makes is this?
Judge what came out, not what went in. New music is audio, one of the four main kinds of things generative AI makes, along with text, images and code. Text is the tempting answer only because her request was typed.
Question 4 · easy
Kiran planned a week of hostel meals in one chat with a chat assistant. Then he opens a new chat and types only: "Make it cheaper." What usually happens?
Inside one chat, the assistant usually takes earlier messages into account, but a new chat usually starts empty, so "Make it cheaper" leaves it guessing. Typing an Aadhaar number never helps and is never safe in a chat.
Question 5 · easy
Sameer asks a chat assistant to explain a bank form in simple English. Which of these should he never type into the chat?
Never type a password, a one-time password, your Aadhaar number, or bank or card details into a chat. Saying who you are, like a commerce graduate, and what you need, like simple English, is habit one and is safe.
Question 6 · easy
Ravi runs a kirana shop in Berhampur and wants a notice for his shop door. Which first message follows habit one, say who you are and what you need, best?
Habit one means one line about your work and what you need, so the reply can suit you. A single word like "Notice" leaves the assistant guessing. Saying who you are does not mean your name, phone number or address, which you should keep out of a chat.
Question 7 · medium
Nisha takes a photo of a handwritten Odia notice and gives it to a chat assistant. The assistant first reads the words in the photo, then writes a new three-line summary in English. Which statement is right?
Many apps do both, so ask the test of each step. Reading the words takes out what was already in the photo, which is recognising; the summary is new text, which is generating. Calling a step generating just because text came out of it is the common trap.
Question 8 · medium
A school app called Answer Maker reads a student's handwritten answer sheet and puts a tick or a cross next to each answer. Is this feature recognising or generating?
Judge the output, not the app's name. A tick or a cross only checks answers the student already wrote, which is recognising. The name "Answer Maker" misleads, and "because it works on paper" is the right label for the wrong reason: the test is whether the output is new.
Question 9 · medium
Kiran types: "I run a kirana shop. Make a price list: 3 kg rice at ₹48 a kg, 2 litres oil at ₹150 a litre." The reply says: "Rice: ₹144. Oil: ₹200." What should he send next?
Replies can be wrong, so check what matters: 3 × ₹48 = ₹144 is right, but 2 × ₹150 = ₹300, not ₹200. On the same topic, correct it plainly in the same chat, saying which part is wrong and what is right. "Thanks, I will print it" trusts it unchecked, "This is wrong. Try again" does not say which line, and a new chat usually starts empty.
Question 10 · medium
Asha used one chat for her son's science project. Now she opens a new chat for a different job and types: "Like before, keep it simple. Write the complaint." What is the problem with this message?
A new topic belongs in a new chat, but a new chat usually starts empty, so "like before" means nothing to it: say who you are and what you need again. Going back to the project chat mixes topics, and a mobile number is personal detail that does not belong in a chat.
Question 11 · medium
Priya types: "I am in Class 8. Explain how plants make food." The reply is correct but long and full of hard words. What should she send next?
The reply was right, only too hard, so she stays in the same chat and follows up, saying what she needs: simpler words and five lines. "Wrong. Do it again." calls a right reply wrong, a new chat usually starts empty so "simpler" has nothing to work on, and repeating the question gets much the same answer.
Question 12 · medium
Sameer asks a chat assistant for a four-line poem about Chilika lake. He opens a new chat, asks again, and gets a different poem. Which explanation matches what this course has said so far?
A large language model is a program that learned patterns from a huge amount of text and uses them to write new text, so each poem is newly written. The tempting picture of a store of ready answers copied out is wrong, his message held no poem to read back, and written rules for every question are the other kind of program.
Question 13 · hard
Meena, a teacher in Cuttack, tries four features on one chapter of a geography book. Which one is generating?
Ask the test: was the output already in what went in? The new quiz questions did not exist before, so they are generated text. The "AI Made Summary" is the trap: its name says made and text comes out, but every highlighted sentence was already in the chapter, so it only picks out what is there. Reading a printed code and sorting pictures are recognising too.
Question 14 · hard
Arjun, a Class 10 student, has used one chat for two weeks: maths doubts, a birthday message, a cricket question, and now biology. The biology replies keep bringing in maths words. What should he do?
A new topic needs a new chat, because old messages can pull new replies off track, and a new chat usually starts empty, so he says who he is and what he needs again. One long chat for everything is the habit causing the trouble, "Wrong, try again" gives no detail, and a single word leaves the new chat guessing.
Question 15 · hard
Nisha asks a chat assistant for the last date to send in a state job exam form. The reply gives an exact date and sounds very sure. What should she do before she relies on it?
Replies can be wrong even when they sound sure, so a date that matters is checked against the official notice before she uses it. Two replies that agree can still both be wrong, and calling a reply wrong without knowing it is wrong can swap a right answer for a wrong one.
Your answers stay in this browser tab. Nothing is sent or saved.