
AI: Myth or Fact?
Explore what AI can do in daily life—and where popular claims go too far.
Preview
Questions with solutions-
1. 正しい答えを選んでください。
You speak into your phone, and your words appear on the screen. What can AI speech recognition commonly do?
- Turn spoken words into text, although mistakes can occur.
- Recognize only sentences recorded during its training.
- Capture every word correctly, even when speech is unclear.
Consider both the useful ability and the effect of background noise.
Show solution
- Turn spoken words into text, although mistakes can occur.
Speech recognition can handle new sentences, but noise, accents, and overlapping voices can affect accuracy.
-
2. 正しい答えを選んでください。
A streaming app recommends a show you enjoy. How does AI commonly produce recommendations like this?
- It knows exactly what you are thinking at that moment.
- It uses patterns in viewing behavior to predict your interests.
- It relies entirely on interests you have explicitly listed.
Your activity can provide clues even when you do not state your preferences.
Show solution
- It uses patterns in viewing behavior to predict your interests.
Recommendations may use what you watch, skip, or rate, along with patterns from other viewers.
-
3. 正しい答えを選んでください。
You ask an image generator for a purple bicycle floating above a desert. What can this kind of AI do?
- Find only an existing photograph that exactly matches the request.
- Build every result by cutting complete objects out of stored photographs.
- Generate a new image using visual patterns learned during training.
Learning from images is not the same as searching for a finished picture.
Show solution
- Generate a new image using visual patterns learned during training.
Image generators can combine familiar visual patterns in new ways; they are not simply photo-search tools.
-
4. 正しい答えを選んでください。
A travel review contains an idiom rather than a literal description. What can a modern AI translation system do with it?
- Use context to translate its meaning, though it may still misunderstand.
- Translate only the separate words, without using the surrounding sentence.
- Always preserve every cultural meaning without needing to be checked by a person.
Context helps, but it does not guarantee perfect understanding.
Show solution
- Use context to translate its meaning, though it may still misunderstand.
Idioms often need a different expression in another language rather than a word-for-word translation.
-
5. 正しい答えを選んでください。
A chatbot confidently names a book and gives a quotation from it. What is true even if the answer sounds convincing?
- Its confident wording guarantees that the book and quotation are real.
- It may have invented the book or the quotation.
- Giving a quotation means it has checked the original source.
A natural-sounding answer and a verified answer are different things.
Show solution
- It may have invented the book or the quotation.
AI can generate realistic-looking references that do not exist, so important sources need independent checking.
-
6. 正しい答えを選んでください。
A chatbot's training ended last month, and it has no access to live information. What can it reliably know about an event that happened yesterday?
- The latest details, because every chatbot automatically receives news updates.
- The exact outcome, because earlier patterns reveal later events with certainty.
- Nothing, unless someone supplies current information.
Consider where information about a new event would come from.
Show solution
- Nothing, unless someone supplies current information.
Training knowledge is not a live news feed. Some systems gain current information through search tools or material users provide.
-
7. 正しい答えを選んでください。
A company trains a hiring AI on past decisions that unfairly favored one group. What can happen even if the software never explicitly mentions that group?
- It may repeat unfair patterns through other details in the data.
- It becomes fair automatically because computers have no personal opinions.
- Removing the group's name guarantees that all unfair patterns disappear.
Other details can be closely connected to group membership.
Show solution
- It may repeat unfair patterns through other details in the data.
Details such as location can indirectly reflect group membership, allowing unfair patterns to survive.
-
8. 正しい答えを選んでください。
An AI system defeats a champion at chess. What does this achievement establish about its abilities?
- It can handle unfamiliar everyday problems as well as the champion can.
- It has exceptional skill at chess, not necessarily at unrelated tasks.
- It can learn any human skill without further training or changes.
A clear set of game rules is different from the variety of daily life.
Show solution
- It has exceptional skill at chess, not necessarily at unrelated tasks.
Many AI systems are specialized: outstanding performance in one area does not automatically transfer to others.
-
9. 正しい答えを選んでください。
An AI tool has been properly tested for use in detecting signs of a particular disease in medical scans. What can it do in patient care?
- Guarantee that every future case of the disease will be detected.
- Remove the need for medical review whenever a scan result is uncertain.
- Help clinicians spot suspicious patterns and make informed decisions.
A useful medical tool can support a professional without being perfect.
Show solution
- Help clinicians spot suspicious patterns and make informed decisions.
Clinicians can combine AI findings with symptoms, medical history, and other tests that a scan alone may not capture.
-
10. 正しい答えを選んでください。
A driverless taxi service operates in a mapped area but pauses during severe storms. What does this show about self-driving AI?
- Its autonomous driving ability can be limited to certain places and conditions.
- Mapping a road makes weather irrelevant to safe autonomous driving.
- Success in one area guarantees safe operation on all public roads.
Being autonomous does not mean being suitable for every environment.
Show solution
- Its autonomous driving ability can be limited to certain places and conditions.
Weather can affect sensors and road conditions, even when the system already knows the route.
-
11. 正しい答えを選んでください。
An AI-text detector labels a student's essay "AI-written." What is true about this kind of result?
- It proves that at least one sentence was generated by AI.
- It is an estimate that can wrongly flag writing produced by a person.
- It identifies the exact AI model that produced the essay.
Patterns in wording do not provide a complete record of who wrote a text.
Show solution
- It is an estimate that can wrongly flag writing produced by a person.
Human and AI writing can share features, so text detectors can produce both false accusations and missed detections.
-
12. 正しい答えを選んでください。
A chatbot says, "I feel lonely," and gives thoughtful replies. What is true about this behavior?
- A believable description of loneliness confirms that it experiences that feeling.
- Its use of the word "I" proves that it has a human-like sense of self.
- Emotional language alone does not establish that it experiences feelings.
Describing an experience and having that experience are not the same.
Show solution
- Emotional language alone does not establish that it experiences feelings.
Training includes human conversations and stories, so chatbots can reproduce emotional expressions without this proving consciousness.