Talking to Machines
What is abstraction in computing and AI?
a) Adding extra details to make a task more complex
b) Focusing only on the important information
c) Rewriting text to remove errors
d) Using images instead of words
Focusing only on the important information
A large language model (LLM) is:
a) A small program that automates chores
b) A large database used for storage
c) A form of AI trained on massive amounts of text
c) A form of AI trained on massive amounts of text
Context helps an AI model by:
a) Removing unnecessary information
b) Providing additional details to improve accuracy
c) Limiting how much the model can generate
b) Providing additional details to improve accuracy
Prompt engineering involves:
a) Writing very long questions
b) Creating precise prompts to guide AI effectively
c) Removing all user input
b) Creating precise prompts to guide AI effectively
Attribution is the practice of:
a) Selling someone else’s work
b) Giving credit to the original creator
c) Hiding where information came from
b) Giving credit to the original creator
What is artificial intelligence (AI)?
a) A computer that stores large amounts of data
b) A tool that mimics human intelligence and performs tasks
c) A type of social media algorithm
d) A device used only for math calculations
b) A tool that mimics human intelligence and performs tasks
A neural network works by:
a) Running only on physical neurons
b) Using connected nodes, weights, and hidden layers
c) Guessing answers based on a coin flip
b) Using connected nodes, weights, and hidden layers
A contradiction happens when an AI model:
a) Gives two responses that cannot both be true
b) Uses too much training data
c) Shares private information
a) Gives two responses that cannot both be true
Co-creating means:
a) Humans and AI working together
b) AI working alone without humans
c) Humans gathering data without AI
a) Humans and AI working together
Copyright gives creators the right to:
a) Make unlimited copies of all online content
b) Use, distribute, and modify their own original work
c) Share any work freely without permission
b) Use, distribute, and modify their own original work
A prompt is best described as:
a) A visual image generated by AI
b) A question or instruction the user provides
c) A mathematical function in a neural network
d) A mistake made by a computer
b) A question or instruction the user provides
Training an AI model means:
a) Playing games with it
b) Exposing it to lots of data so it learns patterns
c) Deleting old information
b) Exposing it to lots of data so it learns patterns
AI bias occurs when:
a) The model performs tasks too quickly
b) A model favors certain information unfairly
c) The model stops learning
b) A model favors certain information unfairly
A student asks an AI chatbot for bike helmet laws in their city, and the AI generates a law
that doesn't actually exist. What is the MOST likely impact of this hallucination?
a) The student realizes AI is unreliable and stops using it for research.
b) The AI automatically corrects itself when new data becomes available.
c) The student believes the false law and shares it on social media, spreading misinformation.
c) The student believes the false law and shares it on social media, spreading misinformatio
Ethics in technology involves:
a) Choosing the fastest tools
b) Following guidelines for privacy, consent, and safety
c) Saving money on data storage
b) Following guidelines for privacy, consent, and safety
What does probability measure?
a) How often AI needs updating
b) The speed at which a model trains
c) The likelihood that something will occur
c) The likelihood that something will occur
Training data refers to:
a) Only numbers used for calculations
b) Random information unrelated to AI
c) The text, images, or other information used to teach an AI model
c) The text, images, or other information used to teach an AI model
Alignment in AI focuses on:
a) Making the model faster
b) Reducing storage usage
c) Ensuring AI systems match human values and fairness
c) Ensuring AI systems match human values and fairness
What is the BEST strategy to reduce the risk of AI hallucinations when asking for information?
a) Ask open-ended questions so the AI has more freedom to generate creative responses.
b) Be specific in your prompts, check responses for accuracy, and verify facts with external sources.
c)Trust AI-generated content as long as it includes references, even if the references are not clickable.
b) Be specific in your prompts, check responses for accuracy, and verify facts with external sources.
Transparency means:
a) Hiding how decisions are made
b) Deleting training data
c) Openly explaining how systems and processes work
c) Openly explaining how systems and processes work
A multimodal model can:
a) Only understand text
b) Only generate images
c) Process multiple types of input, like text and images
c) Process multiple types of input, like text and images
A hallucination in AI occurs when the model:
a) Gives correct but incomplete information
b) Generates false or misleading information that sounds real
c) Refuses to answer a prompt
b) Generates false or misleading information that sounds real
What is one major weakness of AI compared to human intelligence?
a) AI lacks creativity and true understanding, relying only on patterns and probability.
b) AI makes no mistakes because it processes data with perfect accuracy.
c) AI can reason and think critically just like humans.
a) AI lacks creativity and true understanding, relying only on patterns and probability.
Which best describes the purpose of adding context in an AI prompt?
a) It forces the AI to give longer answers with extra filler details.
b) It helps the AI focus on relevant background information to produce a more tailored response.
c) It automatically overrides any task instructions the user previously gave.
b) It helps the AI focus on relevant background information to produce a more tailored response.
What is one reason AI-generated content may still need human revision?
a) AI's ideas are always wrong.
b) AI doesn't understand audience, emotion, or deeper meaning.
c) AI content is usually unreadable.
b) AI doesn't understand audience, emotion, or deeper meaning
When using AI to get feedback on a digital art project, why might it be beneficial to provide both text and an image?
a) The AI can analyze colors, shapes, and artistic techniques while also considering the artist's intent.
b) The AI will ignore the image and only read the description.
c) Text descriptions are always more useful than images for art critiques.
a) The AI can analyze colors, shapes, and artistic techniques while also considering the artist's intent.
When is AI the BEST tool for solving a problem?
a) When the task requires ethics and emotional reasoning.
b) When the task involves recognizing large-scale patterns in data quickly.
c) When the task requires unique human experience.
b) When the task involves recognizing large-scale patterns in data quickly.
Which scenario best illustrates when an AI contradiction might be harmless?
a) Two AI chatbots in a legal setting each interpret laws differently for a court trial.
b) An AI that provides contradictory medical advice to someone experiencing severe pain.
c) An AI recommending conflicting movie suggestions to a user looking for weekend entertainment
c) An AI recommending conflicting movie suggestions to a user looking for weekend entertainment
What is one reason humans are still essential in creative work with AI?
a) AI can only be used to generate code, not creative content.
b)AI does not understand user input without being explicitly programmed)
c) AI-generated content often lacks personal insight, emotion, and originality.
c) AI-generated content often lacks personal insight, emotion, and originality.
Why might creators want the ability to opt out of having their work used in AI training?
a) To improve the quality of AI outputs.
b) To ensure AI systems remain objective.
c) To maintain control over how their work is used and avoid exploitation.
c) To maintain control over how their work is used and avoid exploitation.
Which of the following is the BEST way to refine this prompt for a more useful AI response?
a) Animals exist. Explain them to me.
b) Tell me everything you know about every animal in the rainforest.
c) Tell me three interesting facts about rainforest animals.
c) Tell me three interesting facts about rainforest animals.
What is one major weakness of AI compared to human intelligence?
a) AI lacks creativity and true understanding, relying only on patterns and probability.
b) AI makes no mistakes because it processes data with perfect accuracy.
c) AI can reason and think critically just like humans.
a) AI lacks creativity and true understanding, relying only on patterns and probability.
Which of the following strategies can help mitigate AI bias?
a) Ensuring training data is more diverse and representative
b) Adding more data without filtering or reviewing it
c) Letting AI self-correct over time without human intervention
a) Ensuring training data is more diverse and representative
Why is it valuable for humans and AI to work together on creative tasks?
a) It makes content completely original with no flaws.
b) AI handles all the work so humans can avoid thinking.
c) AI brings speed and pattern recognition, while humans bring meaning and creativity.
c) AI brings speed and pattern recognition, while humans bring meaning and creativity.
Which example best shows a positive environmental strategy for managing AI's energy
use?
a) Training large AI models during peak energy hours.
b)Relocating data centers to areas with high pollution.
c) Powering AI data centers using renewable energy sources.
c) Powering AI data centers using renewable energy sources.