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
A program is made from written rules. What does that mean?
In a rules program, a person writes every step in advance and the program follows them, so the same input always gives the same result. Finding the steps from many examples is the other kind, the learned kind. Being fast and exact does not make a program AI, and neither kind understands like a person.
Question 2 · easy
When people say AI today, what do they almost always mean?
Today AI almost always means the learned kind: a program that found patterns in many examples instead of following steps a person wrote. A fast or clever app can be plain rules, being online does not make an app AI, and there is no person inside the phone.
Question 3 · easy
Asha's kirana shop billing app adds up the price of each item she scans and prints the total. Which kind of program does the adding?
Adding prices follows steps a person can write down in advance, so it is the rules kind, and the same items always give the same total. Being fast and exact feels clever, but clever is not the same as learned, and running on a computer says nothing about how a program was made.
Question 4 · easy
Priya's music app suggests songs she has never heard, and she often likes them. Nobody wrote a list of which song suits which listener. Which kind of program is the suggestion feature?
No person could write a rule for every listener and every song, so a suggestion feature like this usually learns from what many listeners played: the examples kind. A person picking songs for each listener could never cover everyone, and being online does not decide the kind.
Question 5 · easy
Kiran asks the voice assistant on her phone a question and gets a helpful answer. She says, "There must be a person inside who understands me." What is the right view?
The helpful answer comes from a program that learned patterns from data. There is no person inside the phone, and the program does not understand like a friend; it matches patterns. Calling it a calculator with a voice misses that it learned from examples.
Question 6 · easy
Sameer's translation app usually turns his Odia sentences into good English. One day it gives a sentence with the wrong meaning. What does this most likely show?
A learned program works from patterns, so it handles new sentences but now and then gets one wrong; check the ones that matter. Thinking a learned program is never wrong is the trap, nobody wrote a rule for every sentence, and a program does not get tired.
Question 7 · medium
A college app has four features. Which one learned from examples?
Handwriting differs from person to person, so no one could write a step for every way of writing a letter; such features usually learn from many samples of handwriting. The attendance warning, the ₹2 fine and the timetable each follow a rule a person can write in one line, even if a warning feels like a judgement.
Question 8 · medium
Meena's shopping app has four features. Which one follows written rules only?
A limit of 10 per order is one line a person can write in advance, so it is the rules kind, even though refusing an order feels like a decision. Suggestions from many shoppers, catching fake reviews in new wording and matching a photo all handle cases no one listed, so they usually learn from examples.
Question 9 · medium
Nisha is not sure whether a feature on her phone is the rules kind or the examples kind. Which one question sorts it best?
Ask how it was made: if a person could write down every step in advance, it is the rules kind; if no one could list every case and it learned from many examples, it is the examples kind. Speed, being online and a friendly voice say nothing about how a program was made.
Question 10 · medium
Ravi's bank app shows his balance, adds interest at a fixed rate each month, and sends an alert when a payment looks unusual for him. Ravi says, "The whole app is AI." What is the best reply?
Many apps mix both kinds, so sort each feature on its own. Showing the balance and adding interest at a fixed rate follow written steps, while spotting an unusual payment means handling tricks no one listed, so it usually learns from many past payments. Calling the whole app AI, or the whole app not AI, skips the sorting, and being online decides nothing.
Question 11 · medium
Arjun hums a tune into an app and it names the song, even though he hums off-key and has never hummed that tune into it before. Sometimes it names the wrong song. Which kind of program is it, and what gives it away?
Two signs point to learning: it copes with hums it has never heard, and now and then it names the wrong song. No one could store a hum of every song by every person, a wrong guess is normal for a program that matches patterns rather than a sign of a mistyped rule, and speed says nothing about how a program was made.
Question 12 · medium
A district office in Balasore gets thousands of complaint SMS every week. People write in their own words, in Odia, English or a mix. The office wants each SMS sent to the right department: water, roads or electricity. Why would a learned program usually suit this better than written rules?
When people write in endless new ways, no person could write a rule for every message, so a program that learned from many messages staff had already sorted usually does better. It will still send some messages to the wrong place, since learned programs slip now and then, and it matches patterns rather than understanding like an officer. Rules work fine on SMS whenever every case can be written down.
Question 13 · hard
Arjun, a clerk in Cuttack, uses office software with five features: - 1: puts today's date on every letter - 2: blocks a leave request longer than 30 days until a manager approves it - 3: flags a sentence in a draft letter that may sound rude, even with words it has never seen - 4: totals the month's travel claims - 5: groups photos of old files by what they show (receipts, letters, maps), even photos it has never seen
Which features learned from examples?
Features 3 and 5 handle cases no one listed, so they usually learn from many letters and photos that people had already sorted. Blocking leave over 30 days feels like a judgement but is a one-line rule, totalling claims is just adding, and the date is a plain rule. Picking "3 only" forgets that grouping photos by what they show also had to be learned.
Question 14 · hard
Ravi, a farmer near Bargarh, uses a farming app with three features: - converts his land from acres to hectares - shows the crop prices the local mandi posted today - looks at a photo of a paddy leaf and names the likely disease
Which answer can be wrong now and then even when the app works as it was made, so he should check it before buying medicine?
Naming a disease from a photo is learned from many leaf photos, so it works from patterns and now and then is wrong; check what matters before spending money. The conversion and the price list follow written steps and give the same result every time, so one learned feature does not make every answer shaky, and no learned program is always right.
Question 15 · hard
A website's help chat answers only when a message contains certain words. A person typed every reply and the words that trigger it: for example, "refund" gets the refund steps. Any other message gets "Sorry, I did not understand." The advert calls it AI-powered. By the meaning of AI most people use today, which is right?
Ask how it was made: a person wrote every trigger word and every reply, so it is the rules kind, and a new kind of message gets nothing useful. Some older systems built from written rules were called AI too, but today AI usually means a program that learned patterns from examples. Chatting in sentences, a label in an advert or having served many customers does not make it learned.
Your answers stay in this browser tab. Nothing is sent or saved.