by BlastRoom
Identify narrow AI, recognize AI in products, and explain the Turing test.
#computer_science

40 questions
One question set, multiple game modes. Tap any game to see it in action. Not sure which to pick?
Answer to place stones and capture territory.
Outrun the snow together.
Earn blasts, then hunt the other team.
Answer to earn energy, then climb.

The class blasts through walls together.
Bid on mystery lots. Most money wins.
Sell tours. Nice, standard, or a total scam.
Answer, restock, and serve the rush.
More correct answers pulls the rope your way.
Answer, then shoot down the alien wave.
Answer, then tap to leap the frozen ridge.
A system learns from thousands of labeled X-rays to flag images that may show a broken bone. It then flags new X-rays. Which description fits it?
A school bell rings at 8:00 each morning because its timer was programmed with that time. No data was used to change its behavior. How should this system be classified?
A research system can learn a new field from examples, transfer what it learns among unrelated tasks, and reason across those tasks at a humanlike level. What category is this description aiming at?
A music service trains a model on listeners' past plays and skips, then predicts songs one listener may enjoy. Which description is most accurate?
At a warehouse door, a scanner compares each badge number with a preapproved list and unlocks only for a matching number. It does not learn from past entries. What is it?
A translation tool was trained on pairs of sentences in different languages and produces translations for sentences it has not seen before. How should it be classified?
A thermostat turns the heater on when the temperature is below a number chosen by its owner and turns it off otherwise. It never changes that rule. Which description fits?
A game opponent was trained through many practice games to choose moves that increase its chance of winning one board game. What is it?
A photo kiosk places prints into folders by matching the date in each file name to a calendar table. Its maker wrote all of the matching steps in advance. Which classification is best?
A model learns from labeled sound recordings to distinguish a smoke alarm from a doorbell and classifies new recordings. Which description fits?
A phone keyboard suggests the next word by using a model trained on patterns in many text messages. Which part is doing the AI work?
A camera app brightens a photo when its fixed exposure setting is too dark, but it also groups faces by using a model trained on many facial images. Which part uses AI?
A map app calculates a route with fixed road data, then estimates which route will be fastest using a model trained on past traffic patterns. What is the AI subsystem?
An email service puts messages into a spam folder after a model trained on messages labeled spam or not spam scores them. Which component is doing AI work?
A streaming app loads video frames from a server and also chooses a home-page order based on a model trained on viewing histories. Which part uses AI?
A fitness watch counts steps from motion-sensor readings and offers a sleep estimate using a model trained on labeled sleep data. Which feature is the AI feature?
A shopping website stores product photos, processes payments, and shows a customer products predicted from that customer's browsing history. What is doing the AI work?
A voice assistant records sound, sends it across a network, and uses a trained model to turn spoken words into text. Which component is AI?
In Alan Turing's 1950 imitation game, a judge communicates by typed messages with hidden participants. What is the central question?
A hidden machine and a hidden person each exchange typed messages with a judge. The judge identifies the machine correctly in nearly every trial. What does this result mean under the Turing test?
Which detail is intentionally removed in the Turing test setup so a judge focuses on conversation rather than appearance?
Why is a machine's quick response time not enough by itself to pass the Turing test?
A chatbot gives fluent replies in a short typed conversation, and a judge mistakes it for the hidden human. What does that outcome show according to the test?
Which statement best describes what the Turing test does NOT directly measure?
A loan application system adds points for conditions written by its designers, such as “if income is below this amount, add 2 points.” It was not trained on past applications. What kind of system is it?
A plant app adjusts internal numbers while comparing its guesses with thousands of labeled plant photos, reducing errors over many rounds. What evidence shows it uses machine learning?
A museum kiosk answers “open” or “closed” by checking the current time against hours entered by staff. Which change would turn this fixed rule system into a learned system?
A spelling tool was trained on many sentences labeled with corrected spellings and predicts likely corrections for new typing. Why is it not merely a hand-written rule system?
A parking gate camera reads each license plate by using a model trained on thousands of labeled plate photos. The gate then opens only if the plate exactly matches an entry in an approved list. Which part uses AI?
Two systems sort fruit photos. System A uses a designer's rules about color and shape. System B was trained on labeled fruit photos. Which system can update its learned behavior by being retrained on more examples?
A student says, “Because an AI system can identify birds from photos, it can also diagnose any illness without new training.” What is the best response?
A language model writes a confident paragraph containing an invented source citation. What can correctly be concluded?
A student uses a generative AI tool to brainstorm essay topics. Which action best supports academic honesty when using an idea in the final assignment?
A model predicts that an applicant will repay a loan, but its training data includes too few examples from some communities. What risk does this create?
A navigation model predicts traffic using years of data. A new bridge opens and changes travel patterns. Why might its predictions need updating?
What does the term artificial intelligence mean in this course?
What is machine learning?
In machine learning, what is training data?
In an image-classification project, what is a label?
What is generative AI designed to do?
Crack the passcode before the zombies get in.
Answer, then swat bugs for bonus points.
Race through questions at your own speed.
Same question on every screen. Fastest wins.

Pick a tile, answer, hit bonuses and steals.
AI Bias and Fairness
25 questions